Authors: Zahra Chaghazardi (University of Surrey), Saber Fallah (University of Surrey), Alireza Tamaddoni-Nezhad (University of Surrey)
Traffic sign detection is a critical task in the operation of Autonomous Vehicles (AV), as it ensures the safety of all road users. Current DNN-based sign classification systems rely on pixel-level features to detect traffic signs and can be susceptible to adversarial attacks. These attacks involve small, imperceptible changes to a sign that can cause traditional classifiers to misidentify the sign. We propose an Inductive Logic Programming (ILP) based approach for stop sign detection in AVs to address this issue. This method utilises high-level features of a sign, such as its shape, colour, and text, to detect categories of traffic signs. This approach is more robust against adversarial attacks, as it mimics human-like perception and is less susceptible to the limitations of current DNN classifiers. We consider two adversarial attacking methods to evaluate our approach: Robust Physical Perturbation (PR2) and Adversarial Camouflage (AdvCam). These attacks are able to deceive DNN classifiers, causing them to misidentify stop signs as other signs with high confidence. The results show that the proposed ILP-based technique is able to correctly identify all targeted stop signs, even in the presence of PR2 and ADvCam attacks. The proposed learning method is also efficient as it requires minimal training data. Moreover, it is fully explainable, making it possible to debug AVs.
Authors: Bikram Koirala, Behnood Rasti, Zakaria Bnoulkacem, Andrea de Lima Ribeiro, Yuleika Madriz, Erik Herrmann, Arthur Gestels, Thomas De Kerf, Sandra Lorenz, Margret Fuchs, Koen Janssens, Gunther Steenackers, Richard Gloaguen, Paul Scheunders
Optical hyperspectral cameras capture the spectral reflectance of materials. Since many materials behave as heterogeneous intimate mixtures with which each photon interacts differently, the relationship between spectral reflectance and material composition is very complex. Quantitative validation of spectral unmixing algorithms requires high-quality ground truth fractional abundance data, which are very difficult to obtain. In this work, we generated a comprehensive laboratory ground truth dataset of intimately mixed mineral powders. For this, five clay powders (Kaolin, Roof clay, Red clay, mixed clay, and Calcium hydroxide) were mixed homogeneously to prepare 325 samples of 60 binary, 150 ternary, 100 quaternary, and 15 quinary mixtures. Thirteen different hyperspectral sensors have been used to acquire the reflectance spectra of these mixtures in the visible, near, short, mid, and long-wavelength infrared regions (350-15385) nm. {\color{black} Overlaps in wavelength regions due to the operational ranges of each sensor} and variations in acquisition conditions {\color{black} resulted in} a large amount of spectral variability. Ground truth composition is given by construction, but to verify that the generated samples are sufficiently homogeneous, XRD and XRF elemental analysis is performed. We believe these data will be beneficial for validating advanced methods for nonlinear unmixing and material composition estimation, including studying spectral variability and training supervised unmixing approaches. The datasets can be downloaded from the following link: https://github.com/VisionlabUA/Multisensor_datasets.
Authors: Tuan Dinh Nguyen, Keisuke Hihara, Tung Cao Hoang, Yumeka Utada, Akihiko Torii, Naoki Izumi, Nguyen Thanh Thuy, Long Quoc Tran
Understanding customer behavior in retail stores plays a crucial role in improving customer satisfaction by adding personalized value to services. Behavior analysis reveals both general and detailed patterns in the interaction of customers with a store items and other people, providing store managers with insight into customer preferences. Several solutions aim to utilize this data by recognizing specific behaviors through statistical visualization. However, current approaches are limited to the analysis of small customer behavior sets, utilizing conventional methods to detect behaviors. They do not use deep learning techniques such as deep neural networks, which are powerful methods in the field of computer vision. Furthermore, these methods provide limited figures when visualizing the behavioral data acquired by the system. In this study, we propose a framework that includes three primary parts: mathematical modeling of customer behaviors, behavior analysis using an efficient deep learning based system, and individual and group behavior visualization. Each module and the entire system were validated using data from actual situations in a retail store.
Authors: Hengyue Liu, Bir Bhanu
Scene Graph Generation (SGG) has achieved significant progress recently. However, most previous works rely heavily on fixed-size entity representations based on bounding box proposals, anchors, or learnable queries. As each representation's cardinality has different trade-offs between performance and computation overhead, extracting highly representative features efficiently and dynamically is both challenging and crucial for SGG. In this work, a novel architecture called RepSGG is proposed to address the aforementioned challenges, formulating a subject as queries, an object as keys, and their relationship as the maximum attention weight between pairwise queries and keys. With more fine-grained and flexible representation power for entities and relationships, RepSGG learns to sample semantically discriminative and representative points for relationship inference. Moreover, the long-tailed distribution also poses a significant challenge for generalization of SGG. A run-time performance-guided logit adjustment (PGLA) strategy is proposed such that the relationship logits are modified via affine transformations based on run-time performance during training. This strategy encourages a more balanced performance between dominant and rare classes. Experimental results show that RepSGG achieves the state-of-the-art or comparable performance on the Visual Genome and Open Images V6 datasets with fast inference speed, demonstrating the efficacy and efficiency of the proposed methods.
Authors: Nikolai Körber, Eduard Kromer, Andreas Siebert, Sascha Hauke, Daniel Mueller-Gritschneder
We introduce EGIC, a novel generative image compression method that allows traversing the distortion-perception curve efficiently from a single model. Specifically, we propose an implicitly encoded variant of image interpolation that predicts the residual between a MSE-optimized and GAN-optimized decoder output. On the receiver side, the user can then control the impact of the residual on the GAN-based reconstruction. Together with improved GAN-based building blocks, EGIC outperforms a wide-variety of perception-oriented and distortion-oriented baselines, including HiFiC, MRIC and DIRAC, while performing almost on par with VTM-20.0 on the distortion end. EGIC is simple to implement, very lightweight (e.g. 0.18x model parameters compared to HiFiC) and provides excellent interpolation characteristics, which makes it a promising candidate for practical applications targeting the low bit range.
Authors: Mohammed Leo, Kurban Ubul, ShengJie Cheng, Michael Ma
In recent years, Video Object Segmentation (VOS) has emerged as a complementary method to Video Object Tracking (VOT). VOS focuses on classifying all the pixels around the target, allowing for precise shape labeling, while VOT primarily focuses on the approximate region where the target might be. However, traditional segmentation modules usually classify pixels frame by frame, disregarding information between adjacent frames.
In this paper, we propose a new algorithm that addresses this limitation by analyzing the motion pattern using the inherent tensor structure. The tensor structure, obtained through Tucker2 tensor decomposition, proves to be effective in describing the target's motion. By incorporating this information, we achieved competitive results on Four benchmarks LaSOT\cite{fan2019lasot}, AVisT\cite{noman2022avist}, OTB100\cite{7001050}, and GOT-10k\cite{huang2019got} LaSOT\cite{fan2019lasot} with SOTA. Furthermore, the proposed tracker is capable of real-time operation, adding value to its practical application.
Authors: Nikhil Sontakke, Sejal Utekar, Shivansh Rastogi, Shriraj Sonawane
Due to the widespread use of smartphones with high-quality digital cameras and easy access to a wide range of software apps for recording, editing, and sharing videos and images, as well as the deep learning AI platforms, a new phenomenon of 'faking' videos has emerged. Deepfake algorithms can create fake images and videos that are virtually indistinguishable from authentic ones. Therefore, technologies that can detect and assess the integrity of digital visual media are crucial. Deepfakes, also known as deep learning-based fake videos, have become a major concern in recent years due to their ability to manipulate and alter images and videos in a way that is virtually indistinguishable from the original. These deepfake videos can be used for malicious purposes such as spreading misinformation, impersonating individuals, and creating fake news. Deepfake detection technologies use various approaches such as facial recognition, motion analysis, and audio-visual synchronization to identify and flag fake videos. However, the rapid advancement of deepfake technologies has made it increasingly difficult to detect these videos with high accuracy. In this paper, we aim to provide a comprehensive review of the current state of deepfake creation and detection technologies. We examine the various deep learning-based approaches used for creating deepfakes, as well as the techniques used for detecting them. Additionally, we analyze the limitations and challenges of current deepfake detection methods and discuss future research directions in this field. Overall, the paper highlights the importance of continued research and development in deepfake detection technologies in order to combat the negative impact of deepfakes on society and ensure the integrity of digital visual media.
Authors: Yunjie Chen, Marius Staring, Olaf M. Neve, Stephan R. Romeijn, Erik F. Hensen, Berit M. Verbist, Jelmer M. Wolterink, Qian Tao
Multi-sequence magnetic resonance imaging (MRI) has found wide applications in both modern clinical studies and deep learning research. However, in clinical practice, it frequently occurs that one or more of the MRI sequences are missing due to different image acquisition protocols or contrast agent contraindications of patients, limiting the utilization of deep learning models trained on multi-sequence data. One promising approach is to leverage generative models to synthesize the missing sequences, which can serve as a surrogate acquisition. State-of-the-art methods tackling this problem are based on convolutional neural networks (CNN) which usually suffer from spectral biases, resulting in poor reconstruction of high-frequency fine details. In this paper, we propose Conditional Neural fields with Shift modulation (CoNeS), a model that takes voxel coordinates as input and learns a representation of the target images for multi-sequence MRI translation. The proposed model uses a multi-layer perceptron (MLP) instead of a CNN as the decoder for pixel-to-pixel mapping. Hence, each target image is represented as a neural field that is conditioned on the source image via shift modulation with a learned latent code. Experiments on BraTS 2018 and an in-house clinical dataset of vestibular schwannoma patients showed that the proposed method outperformed state-of-the-art methods for multi-sequence MRI translation both visually and quantitatively. Moreover, we conducted spectral analysis, showing that CoNeS was able to overcome the spectral bias issue common in conventional CNN models. To further evaluate the usage of synthesized images in clinical downstream tasks, we tested a segmentation network using the synthesized images at inference.
Authors: Nhat-Tan Bui, Dinh-Hieu Hoang, Quang-Thuc Nguyen, Minh-Triet Tran, Ngan Le
Efficient polyp segmentation in healthcare plays a critical role in enabling early diagnosis of colorectal cancer. However, the segmentation of polyps presents numerous challenges, including the intricate distribution of backgrounds, variations in polyp sizes and shapes, and indistinct boundaries. Defining the boundary between the foreground (i.e. polyp itself) and the background (surrounding tissue) is difficult. To mitigate these challenges, we propose Multi-Scale Edge-Guided Attention Network (MEGANet) tailored specifically for polyp segmentation within colonoscopy images. This network draws inspiration from the fusion of a classical edge detection technique with an attention mechanism. By combining these techniques, MEGANet effectively preserves high-frequency information, notably edges and boundaries, which tend to erode as neural networks deepen. MEGANet is designed as an end-to-end framework, encompassing three key modules: an encoder, which is responsible for capturing and abstracting the features from the input image, a decoder, which focuses on salient features, and the Edge-Guided Attention module (EGA) that employs the Laplacian Operator to accentuate polyp boundaries. Extensive experiments, both qualitative and quantitative, on five benchmark datasets, demonstrate that our EGANet outperforms other existing SOTA methods under six evaluation metrics. Our code is available at \url{https://github.com/DinhHieuHoang/MEGANet}
Authors: Mengliang Zhang, Xinyue Hu, Lin Gu, Liangchen Liu, Kazuma Kobayashi, Tatsuya Harada, Ronald M. Summers, Yingying Zhu
Patients undergoing chest X-rays (CXR) often endure multiple lung diseases. When evaluating a patient's condition, due to the complex pathologies, subtle texture changes of different lung lesions in images, and patient condition differences, radiologists may make uncertain even when they have experienced long-term clinical training and professional guidance, which makes much noise in extracting disease labels based on CXR reports. In this paper, we re-extract disease labels from CXR reports to make them more realistic by considering disease severity and uncertainty in classification. Our contributions are as follows: 1. We re-extracted the disease labels with severity and uncertainty by a rule-based approach with keywords discussed with clinical experts. 2. To further improve the explainability of chest X-ray diagnosis, we designed a multi-relationship graph learning method with an expert uncertainty-aware loss function. 3. Our multi-relationship graph learning method can also interpret the disease classification results. Our experimental results show that models considering disease severity and uncertainty outperform previous state-of-the-art methods.
Authors: Nivetha Jayakumar, Tonmoy Hossain, Miaomiao Zhang
3D image reconstruction from a limited number of 2D images has been a long-standing challenge in computer vision and image analysis. While deep learning-based approaches have achieved impressive performance in this area, existing deep networks often fail to effectively utilize the shape structures of objects presented in images. As a result, the topology of reconstructed objects may not be well preserved, leading to the presence of artifacts such as discontinuities, holes, or mismatched connections between different parts. In this paper, we propose a shape-aware network based on diffusion models for 3D image reconstruction, named SADIR, to address these issues. In contrast to previous methods that primarily rely on spatial correlations of image intensities for 3D reconstruction, our model leverages shape priors learned from the training data to guide the reconstruction process. To achieve this, we develop a joint learning network that simultaneously learns a mean shape under deformation models. Each reconstructed image is then considered as a deformed variant of the mean shape. We validate our model, SADIR, on both brain and cardiac magnetic resonance images (MRIs). Experimental results show that our method outperforms the baselines with lower reconstruction error and better preservation of the shape structure of objects within the images.
Authors: Jiayan Teng, Wendi Zheng, Ming Ding, Wenyi Hong, Jianqiao Wangni, Zhuoyi Yang, Jie Tang
Diffusion models achieved great success in image synthesis, but still face challenges in high-resolution generation. Through the lens of discrete cosine transformation, we find the main reason is that \emph{the same noise level on a higher resolution results in a higher Signal-to-Noise Ratio in the frequency domain}. In this work, we present Relay Diffusion Model (RDM), which transfers a low-resolution image or noise into an equivalent high-resolution one for diffusion model via blurring diffusion and block noise. Therefore, the diffusion process can continue seamlessly in any new resolution or model without restarting from pure noise or low-resolution conditioning. RDM achieves state-of-the-art FID on CelebA-HQ and sFID on ImageNet 256$\times$256, surpassing previous works such as ADM, LDM and DiT by a large margin. All the codes and checkpoints are open-sourced at \url{https://github.com/THUDM/RelayDiffusion}.
Authors: Li Fan, Jeova Farias Sales Rocha Neto
The analysis of Synthetic Aperture Radar (SAR) imagery is an important step in remote sensing applications, and it is a challenging problem due to its inherent speckle noise. One typical solution is to model the data using the $G_I^0$ distribution and extract its roughness information, which in turn can be used in posterior imaging tasks, such as segmentation, classification and interpretation. This leads to the need of quick and reliable estimation of the roughness parameter from SAR data, especially with high resolution images. Unfortunately, traditional parameter estimation procedures are slow and prone to estimation failures. In this work, we proposed a neural network-based estimation framework that first learns how to predict underlying parameters of $G_I^0$ samples and then can be used to estimate the roughness of unseen data. We show that this approach leads to an estimator that is quicker, yields less estimation error and is less prone to failures than the traditional estimation procedures for this problem, even when we use a simple network. More importantly, we show that this same methodology can be generalized to handle image inputs and, even if trained on purely synthetic data for a few seconds, is able to perform real time pixel-wise roughness estimation for high resolution real SAR imagery.
Authors: Venkata Udaya Sameer, Shilpa Mukhopadhyay, Ruchira Naskar, Ishaan Dali
Source camera identification in digital videos is the problem of associating an unknown digital video with its source device, within a closed set of possible devices. The existing techniques in source detection of digital videos try to find a fingerprint of the actual source in the video in form of PRNU (Photo Response Non--Uniformity), and match it against the SPN (Sensor Pattern Noise) of each possible device. The highest correlation indicates the correct source. We investigate the problem of identifying a video source through a feature based approach using machine learning. In this paper, we present a blind forensic technique of video source authentication and identification, based on feature extraction, feature selection and subsequent source classification. The main aim is to determine whether a claimed source for a video is actually its original source. If not, we identify its original source. Our experimental results prove the efficiency of the proposed method compared to traditional fingerprint based technique.
Joint Embedding Architecture-based self-supervised learning methods have attributed the composition of data augmentations as a crucial factor for their strong representation learning capabilities. While regional dropout strategies have proven to guide models to focus on lesser indicative parts of the objects in supervised methods, it hasn't been adopted by self-supervised methods for generating positive pairs. This is because the regional dropout methods are not suitable for the input sampling process of the self-supervised methodology. Whereas dropping informative pixels from the positive pairs can result in inefficient training, replacing patches of a specific object with a different one can steer the model from maximizing the agreement between different positive pairs. Moreover, joint embedding representation learning methods have not made robustness their primary training outcome. To this end, we propose the ViewMix augmentation policy, specially designed for self-supervised learning, upon generating different views of the same image, patches are cut and pasted from one view to another. By leveraging the different views created by this augmentation strategy, multiple joint embedding-based self-supervised methodologies obtained better localization capability and consistently outperformed their corresponding baseline methods. It is also demonstrated that incorporating ViewMix augmentation policy promotes robustness of the representations in the state-of-the-art methods. Furthermore, our experimentation and analysis of compute times suggest that ViewMix augmentation doesn't introduce any additional overhead compared to other counterparts.
Authors: Priyam Mazumdar, Aiman Soliman, Volodymyr Kindratenko, Luigi Marini, Kenton McHenry
The lack of quality labeled data is one of the main bottlenecks for training Deep Learning models. As the task increases in complexity, there is a higher penalty for overfitting and unstable learning. The typical paradigm employed today is Self-Supervised learning, where the model attempts to learn from a large corpus of unstructured and unlabeled data and then transfer that knowledge to the required task. Some notable examples of self-supervision in other modalities are BERT for Large Language Models, Wav2Vec for Speech Recognition, and the Masked AutoEncoder for Vision, which all utilize Transformers to solve a masked prediction task. GeoAI is uniquely poised to take advantage of the self-supervised methodology due to the decades of data collected, little of which is precisely and dependably annotated. Our goal is to extract building and road segmentations from Digital Elevation Models (DEM) that provide a detailed topography of the earths surface. The proposed architecture is the Masked Autoencoder pre-trained on ImageNet (with the limitation that there is a large domain discrepancy between ImageNet and DEM) with an UperNet Head for decoding segmentations. We tested this model with 450 and 50 training images only, utilizing roughly 5% and 0.5% of the original data respectively. On the building segmentation task, this model obtains an 82.1% Intersection over Union (IoU) with 450 Images and 69.1% IoU with only 50 images. On the more challenging road detection task the model obtains an 82.7% IoU with 450 images and 73.2% IoU with only 50 images. Any hand-labeled dataset made today about the earths surface will be immediately obsolete due to the constantly changing nature of the landscape. This motivates the clear necessity for data-efficient learners that can be used for a wide variety of downstream tasks.
Authors: Joshua R. Waite, Jiale Feng, Riley Tavassoli, Laura Harris, Sin Yong Tan, Subhadeep Chakraborty, Soumik Sarkar
The increasing concern surrounding gun violence in the United States has led to a focus on developing systems to improve public safety. One approach to developing such a system is to detect and track shooters, which would help prevent or mitigate the impact of violent incidents. In this paper, we proposed detecting shooters as a whole, rather than just guns, which would allow for improved tracking robustness, as obscuring the gun would no longer cause the system to lose sight of the threat. However, publicly available data on shooters is much more limited and challenging to create than a gun dataset alone. Therefore, we explore the use of domain randomization and transfer learning to improve the effectiveness of training with synthetic data obtained from Unreal Engine environments. This enables the model to be trained on a wider range of data, increasing its ability to generalize to different situations. Using these techniques with YOLOv8 and Deep OC-SORT, we implemented an initial version of a shooter tracking system capable of running on edge hardware, including both a Raspberry Pi and a Jetson Nano.
Authors: Gabriel Efrain Humpire Mamani, Nikolas Lessmann, Ernst Th. Scholten, Mathias Prokop, Colin Jacobs, Bram van Ginneken
In this study, we introduce a deep learning approach for segmenting kidney parenchyma and kidney abnormalities to support clinicians in identifying and quantifying renal abnormalities such as cysts, lesions, masses, metastases, and primary tumors. Our end-to-end segmentation method was trained on 215 contrast-enhanced thoracic-abdominal CT scans, with half of these scans containing one or more abnormalities.
We began by implementing our own version of the original 3D U-Net network and incorporated four additional components: an end-to-end multi-resolution approach, a set of task-specific data augmentations, a modified loss function using top-$k$, and spatial dropout. Furthermore, we devised a tailored post-processing strategy. Ablation studies demonstrated that each of the four modifications enhanced kidney abnormality segmentation performance, while three out of four improved kidney parenchyma segmentation. Subsequently, we trained the nnUNet framework on our dataset. By ensembling the optimized 3D U-Net and the nnUNet with our specialized post-processing, we achieved marginally superior results.
Our best-performing model attained Dice scores of 0.965 and 0.947 for segmenting kidney parenchyma in two test sets (20 scans without abnormalities and 30 with abnormalities), outperforming an independent human observer who scored 0.944 and 0.925, respectively. In segmenting kidney abnormalities within the 30 test scans containing them, the top-performing method achieved a Dice score of 0.585, while an independent second human observer reached a score of 0.664, suggesting potential for further improvement in computerized methods.
All training data is available to the research community under a CC-BY 4.0 license on https://doi.org/10.5281/zenodo.8014289
Authors: Farahnaz Hosseini, Hossein Ebrahimpour, Samaneh Askari
In this paper, we seek a new method in designing an iris recognition system. In this method, first the Haar wavelet features are extracted from iris images. The advantage of using these features is the high-speed extraction, as well as being unique to each iris. Then the back propagation neural network (BPNN) is used as a classifier. In this system, the BPNN parallel algorithms and their implementation on GPUs have been used by the aid of CUDA in order to speed up the learning process. Finally, the system performance and the speeding outcomes in a way that this algorithm is done in series are presented.
Authors: Yalong Jiang, Changkang Li
Video anomaly detection is a challenging task due to the lack in approaches for representing samples. The visual representations of most existing approaches are limited by short-term sequences of observations which cannot provide enough clues for achieving reasonable detections. In this paper, we propose to completely represent the motion patterns of objects by learning from long-term sequences. Firstly, a Stacked State Machine (SSM) model is proposed to represent the temporal dependencies which are consistent across long-range observations. Then SSM model functions in predicting future states based on past ones, the divergence between the predictions with inherent normal patterns and observed ones determines anomalies which violate normal motion patterns. Extensive experiments are carried out to evaluate the proposed approach on the dataset and existing ones. Improvements over state-of-the-art methods can be observed. Our code is available at https://github.com/AllenYLJiang/Anomaly-Detection-in-Sequences.
Authors: Eulrang Cho, Jooyeon Kim, Hyunwoo J. Kim
Pre-trained vision-language models (VLMs) have shown impressive performance on various downstream tasks by utilizing knowledge learned from large data. In general, the performance of VLMs on target tasks can be further improved by prompt tuning, which adds context to the input image or text. By leveraging data from target tasks, various prompt-tuning methods have been studied in the literature. A key to prompt tuning is the feature space alignment between two modalities via learnable vectors with model parameters fixed. We observed that the alignment becomes more effective when embeddings of each modality are `well-arranged' in the latent space. Inspired by this observation, we proposed distribution-aware prompt tuning (DAPT) for vision-language models, which is simple yet effective. Specifically, the prompts are learned by maximizing inter-dispersion, the distance between classes, as well as minimizing the intra-dispersion measured by the distance between embeddings from the same class. Our extensive experiments on 11 benchmark datasets demonstrate that our method significantly improves generalizability. The code is available at https://github.com/mlvlab/DAPT.
Authors: Zehua Ren, Yongheng Sun, Miaomiao Wang, Yuying Feng, Xianjun Li, Chao Jin, Jian Yang, Chunfeng Lian, Fan Wang
Accurate segmentation of punctate white matter lesions (PWMLs) are fundamental for the timely diagnosis and treatment of related developmental disorders. Automated PWMLs segmentation from infant brain MR images is challenging, considering that the lesions are typically small and low-contrast, and the number of lesions may dramatically change across subjects. Existing learning-based methods directly apply general network architectures to this challenging task, which may fail to capture detailed positional information of PWMLs, potentially leading to severe under-segmentations. In this paper, we propose to leverage the idea of counterfactual reasoning coupled with the auxiliary task of brain tissue segmentation to learn fine-grained positional and morphological representations of PWMLs for accurate localization and segmentation. A simple and easy-to-implement deep-learning framework (i.e., DeepPWML) is accordingly designed. It combines the lesion counterfactual map with the tissue probability map to train a lightweight PWML segmentation network, demonstrating state-of-the-art performance on a real-clinical dataset of infant T1w MR images. The code is available at \href{https://github.com/ladderlab-xjtu/DeepPWML}{https://github.com/ladderlab-xjtu/DeepPWML}.
Authors: Yi Tang, Takafumi Iwaguchi, Hiroshi Kawasaki
In this paper, we present an approach to image enhancement with diffusion model in underwater scenes. Our method adapts conditional denoising diffusion probabilistic models to generate the corresponding enhanced images by using the underwater images and the Gaussian noise as the inputs. Additionally, in order to improve the efficiency of the reverse process in the diffusion model, we adopt two different ways. We firstly propose a lightweight transformer-based denoising network, which can effectively promote the time of network forward per iteration. On the other hand, we introduce a skip sampling strategy to reduce the number of iterations. Besides, based on the skip sampling strategy, we propose two different non-uniform sampling methods for the sequence of the time step, namely piecewise sampling and searching with the evolutionary algorithm. Both of them are effective and can further improve performance by using the same steps against the previous uniform sampling. In the end, we conduct a relative evaluation of the widely used underwater enhancement datasets between the recent state-of-the-art methods and the proposed approach. The experimental results prove that our approach can achieve both competitive performance and high efficiency. Our code is available at \href{mailto:https://github.com/piggy2009/DM_underwater}{\color{blue}{https://github.com/piggy2009/DM\_underwater}}.
Authors: Zhuokai Zhao, Harish Palani, Tianyi Liu, Lena Evans, Ruth Toner
Multimodal models have gained significant success in recent years. Standard multimodal approaches often assume unchanged modalities from training stage to inference stage. In practice, however, many scenarios fail to satisfy such assumptions with missing modalities during inference, leading to limitations on where multimodal models can be applied. While existing methods mitigate the problem through reconstructing the missing modalities, it increases unnecessary computational cost, which could be just as critical, especially for large, deployed systems. To solve the problem from both sides, we propose a novel guidance network that promotes knowledge sharing during training, taking advantage of the multimodal representations to train better single-modality models for inference. Real-life experiment in violence detection shows that our proposed framework trains single-modality models that significantly outperform its traditionally trained counterparts while maintaining the same inference cost.
Authors: Yuan Liu, Cheng Lin, Zijiao Zeng, Xiaoxiao Long, Lingjie Liu, Taku Komura, Wenping Wang
In this paper, we present a novel diffusion model called that generates multiview-consistent images from a single-view image. Using pretrained large-scale 2D diffusion models, recent work Zero123 demonstrates the ability to generate plausible novel views from a single-view image of an object. However, maintaining consistency in geometry and colors for the generated images remains a challenge. To address this issue, we propose a synchronized multiview diffusion model that models the joint probability distribution of multiview images, enabling the generation of multiview-consistent images in a single reverse process. SyncDreamer synchronizes the intermediate states of all the generated images at every step of the reverse process through a 3D-aware feature attention mechanism that correlates the corresponding features across different views. Experiments show that SyncDreamer generates images with high consistency across different views, thus making it well-suited for various 3D generation tasks such as novel-view-synthesis, text-to-3D, and image-to-3D.
Authors: Zhuqiang Lu, Kun Hu, Chaoyue Wang, Lei Bai, Zhiyong Wang
A 360-degree (omni-directional) image provides an all-encompassing spherical view of a scene. Recently, there has been an increasing interest in synthesising 360-degree images from conventional narrow field of view (NFoV) images captured by digital cameras and smartphones, for providing immersive experiences in various scenarios such as virtual reality. Yet, existing methods typically fall short in synthesizing intricate visual details or ensure the generated images align consistently with user-provided prompts. In this study, autoregressive omni-aware generative network (AOG-Net) is proposed for 360-degree image generation by out-painting an incomplete 360-degree image progressively with NFoV and text guidances joinly or individually. This autoregressive scheme not only allows for deriving finer-grained and text-consistent patterns by dynamically generating and adjusting the process but also offers users greater flexibility to edit their conditions throughout the generation process. A global-local conditioning mechanism is devised to comprehensively formulate the outpainting guidance in each autoregressive step. Text guidances, omni-visual cues, NFoV inputs and omni-geometry are encoded and further formulated with cross-attention based transformers into a global stream and a local stream into a conditioned generative backbone model. As AOG-Net is compatible to leverage large-scale models for the conditional encoder and the generative prior, it enables the generation to use extensive open-vocabulary text guidances. Comprehensive experiments on two commonly used 360-degree image datasets for both indoor and outdoor settings demonstrate the state-of-the-art performance of our proposed method. Our code will be made publicly available.
Authors: Nikhil Raghuraman, Adam W. Harley, Leonidas Guibas
Current machine learning methods struggle to solve Bongard problems, which are a type of IQ test that requires deriving an abstract "concept" from a set of positive and negative "support" images, and then classifying whether or not a new query image depicts the key concept. On Bongard-HOI, a benchmark for natural-image Bongard problems, existing methods have only reached 66% accuracy (where chance is 50%). Low accuracy is often attributed to neural nets' lack of ability to find human-like symbolic rules. In this work, we point out that many existing methods are forfeiting accuracy due to a much simpler problem: they do not incorporate information contained in the support set as a whole, and rely instead on information extracted from individual supports. This is a critical issue, because unlike in few-shot learning tasks concerning object classification, the "key concept" in a typical Bongard problem can only be distinguished using multiple positives and multiple negatives. We explore a variety of simple methods to take this cross-image context into account, and demonstrate substantial gains over prior methods, leading to new state-of-the-art performance on Bongard-LOGO (75.3%) and Bongard-HOI (72.45%) and strong performance on the original Bongard problem set (60.84%).
Authors: John Chen, Chen Dun, Anastasios Kyrillidis
Advances in Semi-Supervised Learning (SSL) have almost entirely closed the gap between SSL and Supervised Learning at a fraction of the number of labels. However, recent performance improvements have often come \textit{at the cost of significantly increased training computation}. To address this, we propose Curriculum Batch Size (CBS), \textit{an unlabeled batch size curriculum which exploits the natural training dynamics of deep neural networks.} A small unlabeled batch size is used in the beginning of training and is gradually increased to the end of training. A fixed curriculum is used regardless of dataset, model or number of epochs, and reduced training computations is demonstrated on all settings. We apply CBS, strong labeled augmentation, Curriculum Pseudo Labeling (CPL) \citep{FlexMatch} to FixMatch \citep{FixMatch} and term the new SSL algorithm Fast FixMatch. We perform an ablation study to show that strong labeled augmentation and/or CPL do not significantly reduce training computations, but, in synergy with CBS, they achieve optimal performance. Fast FixMatch also achieves substantially higher data utilization compared to previous state-of-the-art. Fast FixMatch achieves between $2.1\times$ - $3.4\times$ reduced training computations on CIFAR-10 with all but 40, 250 and 4000 labels removed, compared to vanilla FixMatch, while attaining the same cited state-of-the-art error rate \citep{FixMatch}. Similar results are achieved for CIFAR-100, SVHN and STL-10. Finally, Fast MixMatch achieves between $2.6\times$ - $3.3\times$ reduced training computations in federated SSL tasks and online/streaming learning SSL tasks, which further demonstrate the generializbility of Fast MixMatch to different scenarios and tasks.
Authors: Xiangjie Sui, Hanwei Zhu, Xuelin Liu, Yuming Fang, Shiqi Wang, Zhou Wang
Despite substantial efforts dedicated to the design of heuristic models for omnidirectional (i.e., 360$^\circ$) image quality assessment (OIQA), a conspicuous gap remains due to the lack of consideration for the diversity of viewing behaviors that leads to the varying perceptual quality of 360$^\circ$ images. Two critical aspects underline this oversight: the neglect of viewing conditions that significantly sway user gaze patterns and the overreliance on a single viewport sequence from the 360$^\circ$ image for quality inference. To address these issues, we introduce a unique generative scanpath representation (GSR) for effective quality inference of 360$^\circ$ images, which aggregates varied perceptual experiences of multi-hypothesis users under a predefined viewing condition. More specifically, given a viewing condition characterized by the starting point of viewing and exploration time, a set of scanpaths consisting of dynamic visual fixations can be produced using an apt scanpath generator. Following this vein, we use the scanpaths to convert the 360$^\circ$ image into the unique GSR, which provides a global overview of gazed-focused contents derived from scanpaths. As such, the quality inference of the 360$^\circ$ image is swiftly transformed to that of GSR. We then propose an efficient OIQA computational framework by learning the quality maps of GSR. Comprehensive experimental results validate that the predictions of the proposed framework are highly consistent with human perception in the spatiotemporal domain, especially in the challenging context of locally distorted 360$^\circ$ images under varied viewing conditions. The code will be released at https://github.com/xiangjieSui/GSR
Authors: Jiajin Tang, Ge Zheng, Sibei Yang
Referring video object segmentation aims to segment a referent throughout a video sequence according to a natural language expression. It requires aligning the natural language expression with the objects' motions and their dynamic associations at the global video level but segmenting objects at the frame level. To achieve this goal, we propose to simultaneously maintain a global referent token and a sequence of object queries, where the former is responsible for capturing video-level referent according to the language expression, while the latter serves to better locate and segment objects with each frame. Furthermore, to explicitly capture object motions and spatial-temporal cross-modal reasoning over objects, we propose a novel temporal collection-distribution mechanism for interacting between the global referent token and object queries. Specifically, the temporal collection mechanism collects global information for the referent token from object queries to the temporal motions to the language expression. In turn, the temporal distribution first distributes the referent token to the referent sequence across all frames and then performs efficient cross-frame reasoning between the referent sequence and object queries in every frame. Experimental results show that our method outperforms state-of-the-art methods on all benchmarks consistently and significantly.
Authors: Lemeng Wang, Wentao Liu, Weijin Xu, Haoyuan Li, Huihua Yang, Feng Gao
Cerebrovascular disease is one of the major diseases facing the world today. Automatic segmentation of intracranial artery (IA) in digital subtraction angiography (DSA) sequences is an important step in the diagnosis of vascular related diseases and in guiding neurointerventional procedures. While, a single image can only show part of the IA within the contrast medium according to the imaging principle of DSA technology. Therefore, 2D DSA segmentation methods are unable to capture the complete IA information and treatment of cerebrovascular diseases. We propose A timing sequence image segmentation network with U-shape, called TSI-Net, which incorporates a bi-directional ConvGRU module (BCM) in the encoder. The network incorporates a bi-directional ConvGRU module (BCM) in the encoder, which can input variable-length DSA sequences, retain past and future information, segment them into 2D images. In addition, we introduce a sensitive detail branch (SDB) at the end for supervising fine vessels. Experimented on the DSA sequence dataset DIAS, the method performs significantly better than state-of-the-art networks in recent years. In particular, it achieves a Sen evaluation metric of 0.797, which is a 3% improvement compared to other methods.
Authors: Clarence Lee, M Ganesh Kumar, Cheston Tan
State-of-the-art visual grounding models can achieve high detection accuracy, but they are not designed to distinguish between all objects versus only certain objects of interest. In natural language, in order to specify a particular object or set of objects of interest, humans use determiners such as "my", "either" and "those". Determiners, as an important word class, are a type of schema in natural language about the reference or quantity of the noun. Existing grounded referencing datasets place much less emphasis on determiners, compared to other word classes such as nouns, verbs and adjectives. This makes it difficult to develop models that understand the full variety and complexity of object referencing. Thus, we have developed and released the DetermiNet dataset , which comprises 250,000 synthetically generated images and captions based on 25 determiners. The task is to predict bounding boxes to identify objects of interest, constrained by the semantics of the given determiner. We find that current state-of-the-art visual grounding models do not perform well on the dataset, highlighting the limitations of existing models on reference and quantification tasks.
Authors: Nhat-Tan Bui, Dinh-Hieu Hoang, Minh-Triet Tran, Ngan Le
Image segmentation is a critical task in medical image analysis, providing valuable information that helps to make an accurate diagnosis. In recent years, deep learning-based automatic image segmentation methods have achieved outstanding results in medical images. In this paper, inspired by the Segment Anything Model (SAM), a foundation model that has received much attention for its impressive accuracy and powerful generalization ability in 2D still image segmentation, we propose a SAM3D that targets at 3D volumetric medical images and utilizes the pre-trained features from the SAM encoder to capture meaningful representations of input images. Different from other existing SAM-based volumetric segmentation methods that perform the segmentation by dividing the volume into a set of 2D slices, our model takes the whole 3D volume image as input and processes it simply and effectively that avoids training a significant number of parameters. Extensive experiments are conducted on multiple medical image datasets to demonstrate that our network attains competitive results compared with other state-of-the-art methods in 3D medical segmentation tasks while being significantly efficient in terms of parameters.
Authors: Christoph Wies, Lucas Schneide, Sarah Haggenmueller, Tabea-Clara Bucher, Sarah Hobelsberger, Markus V. Heppt, Gerardo Ferrara, Eva I. Krieghoff-Henning, Titus J. Brinker
Pathologists routinely use immunohistochemical (IHC)-stained tissue slides against MelanA in addition to hematoxylin and eosin (H&E)-stained slides to improve their accuracy in diagnosing melanomas. The use of diagnostic Deep Learning (DL)-based support systems for automated examination of tissue morphology and cellular composition has been well studied in standard H&E-stained tissue slides. In contrast, there are few studies that analyze IHC slides using DL. Therefore, we investigated the separate and joint performance of ResNets trained on MelanA and corresponding H&E-stained slides. The MelanA classifier achieved an area under receiver operating characteristics curve (AUROC) of 0.82 and 0.74 on out of distribution (OOD)-datasets, similar to the H&E-based benchmark classification of 0.81 and 0.75, respectively. A combined classifier using MelanA and H&E achieved AUROCs of 0.85 and 0.81 on the OOD datasets. DL MelanA-based assistance systems show the same performance as the benchmark H&E classification and may be improved by multi stain classification to assist pathologists in their clinical routine.
Authors: Karina Ruzaeva, Kishan Govind, Marc Legros, Stefan Sandfeld
Quantitative Transmission Electron Microscopy (TEM) during in-situ straining experiment is able to reveal the motion of dislocations -- linear defects in the crystal lattice of metals. In the domain of materials science, the knowledge about the location and movement of dislocations is important for creating novel materials with superior properties. A long-standing problem, however, is to identify the position and extract the shape of dislocations, which would ultimately help to create a digital twin of such materials. In this work, we quantitatively compare state-of-the-art instance segmentation methods, including Mask R-CNN and YOLOv8. The dislocation masks as the results of the instance segmentation are converted to mathematical lines, enabling quantitative analysis of dislocation length and geometry -- important information for the domain scientist, which we then propose to include as a novel length-aware quality metric for estimating the network performance. Our segmentation pipeline shows a high accuracy suitable for all domain-specific, further post-processing. Additionally, our physics-based metric turns out to perform much more consistently than typically used pixel-wise metrics.
Authors: Teng Hu, Ran Yi, Haokun Zhu, Liang Liu, Jinlong Peng, Yabiao Wang, Chengjie Wang, Lizhuang Ma
Stroke-based rendering aims to recreate an image with a set of strokes. Most existing methods render complex images using an uniform-block-dividing strategy, which leads to boundary inconsistency artifacts. To solve the problem, we propose Compositional Neural Painter, a novel stroke-based rendering framework which dynamically predicts the next painting region based on the current canvas, instead of dividing the image plane uniformly into painting regions. We start from an empty canvas and divide the painting process into several steps. At each step, a compositor network trained with a phasic RL strategy first predicts the next painting region, then a painter network trained with a WGAN discriminator predicts stroke parameters, and a stroke renderer paints the strokes onto the painting region of the current canvas. Moreover, we extend our method to stroke-based style transfer with a novel differentiable distance transform loss, which helps preserve the structure of the input image during stroke-based stylization. Extensive experiments show our model outperforms the existing models in both stroke-based neural painting and stroke-based stylization. Code is available at https://github.com/sjtuplayer/Compositional_Neural_Painter
Authors: Thanh-Huy Nguyen, Quang Hien Kha, Thai Ngoc Toan Truong, Ba Thinh Lam, Ba Hung Ngo, Quang Vinh Dinh, Nguyen Quoc Khanh Le
In recent years, many mammographic image analysis methods have been introduced for improving cancer classification tasks. Two major issues of mammogram classification tasks are leveraging multi-view mammographic information and class-imbalance handling. In the first problem, many multi-view methods have been released for concatenating features of two or more views for the training and inference stage. Having said that, most multi-view existing methods are not explainable in the meaning of feature fusion, and treat many views equally for diagnosing. Our work aims to propose a simple but novel method for enhancing examined view (main view) by leveraging low-level feature information from the auxiliary view (ipsilateral view) before learning the high-level feature that contains the cancerous features. For the second issue, we also propose a simple but novel malignant mammogram synthesis framework for upsampling minor class samples. Our easy-to-implement and no-training framework has eliminated the current limitation of the CutMix algorithm which is unreliable synthesized images with random pasted patches, hard-contour problems, and domain shift problems. Our results on VinDr-Mammo and CMMD datasets show the effectiveness of our two new frameworks for both multi-view training and synthesizing mammographic images, outperforming the previous conventional methods in our experimental settings.
Authors: Lingtong Kong, Boyuan Jiang, Donghao Luo, Wenqing Chu, Ying Tai, Chengjie Wang, Jie Yang
Video frame interpolation is an important low-level vision task, which can increase frame rate for more fluent visual experience. Existing methods have achieved great success by employing advanced motion models and synthesis networks. However, the spatial redundancy when synthesizing the target frame has not been fully explored, that can result in lots of inefficient computation. On the other hand, the computation compression degree in frame interpolation is highly dependent on both texture distribution and scene motion, which demands to understand the spatial-temporal information of each input frame pair for a better compression degree selection. In this work, we propose a novel two-stage frame interpolation framework termed WaveletVFI to address above problems. It first estimates intermediate optical flow with a lightweight motion perception network, and then a wavelet synthesis network uses flow aligned context features to predict multi-scale wavelet coefficients with sparse convolution for efficient target frame reconstruction, where the sparse valid masks that control computation in each scale are determined by a crucial threshold ratio. Instead of setting a fixed value like previous methods, we find that embedding a classifier in the motion perception network to learn a dynamic threshold for each sample can achieve more computation reduction with almost no loss of accuracy. On the common high resolution and animation frame interpolation benchmarks, proposed WaveletVFI can reduce computation up to 40% while maintaining similar accuracy, making it perform more efficiently against other state-of-the-arts. Code is available at https://github.com/ltkong218/WaveletVFI.
Authors: Jiatai Lin, Guoqiang Han, Xuemiao Xu, Changhong Liang, Tien-Tsin Wong, C. L. Philip Chen, Zaiyi Liu, Chu Han
Class activation mapping~(CAM), a visualization technique for interpreting deep learning models, is now commonly used for weakly supervised semantic segmentation~(WSSS) and object localization~(WSOL). It is the weighted aggregation of the feature maps by activating the high class-relevance ones. Current CAM methods achieve it relying on the training outcomes, such as predicted scores~(forward information), gradients~(backward information), etc. However, when with small-scale data, unstable training may lead to less effective model outcomes and generate unreliable weights, finally resulting in incorrect activation and noisy CAM seeds. In this paper, we propose an outcome-agnostic CAM approach, called BroadCAM, for small-scale weakly supervised applications. Since broad learning system (BLS) is independent to the model learning, BroadCAM can avoid the weights being affected by the unreliable model outcomes when with small-scale data. By evaluating BroadCAM on VOC2012 (natural images) and BCSS-WSSS (medical images) for WSSS and OpenImages30k for WSOL, BroadCAM demonstrates superior performance than existing CAM methods with small-scale data (less than 5\%) in different CNN architectures. It also achieves SOTA performance with large-scale training data. Extensive qualitative comparisons are conducted to demonstrate how BroadCAM activates the high class-relevance feature maps and generates reliable CAMs when with small-scale training data.
Authors: Arne Moos
This paper proposes a novel approach for detecting objects using mobile robots in the context of the RoboCup Standard Platform League, with a primary focus on detecting the ball. The challenge lies in detecting a dynamic object in varying lighting conditions and blurred images caused by fast movements. To address this challenge, the paper presents a convolutional neural network architecture designed specifically for computationally constrained robotic platforms. The proposed CNN is trained to achieve high precision classification of single objects in image patches and to determine their precise spatial positions. The paper further integrates Early Exits into the existing high-precision CNN architecture to reduce the computational cost of easily rejectable cases in the background class. The training process involves a composite loss function based on confidence and positional losses with dynamic weighting and data augmentation. The proposed approach achieves a precision of 100% on the validation dataset and a recall of almost 87%, while maintaining an execution time of around 170 $\mu$s per hypotheses. By combining the proposed approach with an Early Exit, a runtime optimization of more than 28%, on average, can be achieved compared to the original CNN. Overall, this paper provides an efficient solution for an enhanced detection of objects, especially the ball, in computationally constrained robotic platforms.
Authors: Sandipan Choudhuri, Suli Adeniye, Arunabha Sen
This work proposes a robust Partial Domain Adaptation (PDA) framework that mitigates the negative transfer problem by incorporating a robust target-supervision strategy. It leverages ensemble learning and includes diverse, complementary label feedback, alleviating the effect of incorrect feedback and promoting pseudo-label refinement. Rather than relying exclusively on first-order moments for distribution alignment, our approach offers explicit objectives to optimize intra-class compactness and inter-class separation with the inferred source prototypes and highly-confident target samples in a domain-invariant fashion. Notably, we ensure source data privacy by eliminating the need to access the source data during the adaptation phase through a priori inference of source prototypes. We conducted a series of comprehensive experiments, including an ablation analysis, covering a range of partial domain adaptation tasks. Comprehensive evaluations on benchmark datasets corroborate our framework's enhanced robustness and generalization, demonstrating its superiority over existing state-of-the-art PDA approaches.
Authors: Tariq M. Khan, Muhammad Arsalan, Shahzaib Iqbal, Imran Razzak, Erik Meijering
Diseases such as diabetic retinopathy and age-related macular degeneration pose a significant risk to vision, highlighting the importance of precise segmentation of retinal vessels for the tracking and diagnosis of progression. However, existing vessel segmentation methods that heavily rely on encoder-decoder structures struggle to capture contextual information about retinal vessel configurations, leading to challenges in reconciling semantic disparities between encoder and decoder features. To address this, we propose a novel feature enhancement segmentation network (FES-Net) that achieves accurate pixel-wise segmentation without requiring additional image enhancement steps. FES-Net directly processes the input image and utilizes four prompt convolutional blocks (PCBs) during downsampling, complemented by a shallow upsampling approach to generate a binary mask for each class. We evaluate the performance of FES-Net on four publicly available state-of-the-art datasets: DRIVE, STARE, CHASE, and HRF. The evaluation results clearly demonstrate the superior performance of FES-Net compared to other competitive approaches documented in the existing literature.
Authors: Chenjie Zhang, Pengcheng Jiao
You Only Look Once (YOLO) algorithm is a representative target detection algorithm emerging in 2016, which is known for its balance of computing speed and accuracy, and now plays an important role in various fields of human production and life. However, there are still many limitations in the application of YOLO algorithm in underwater environments due to problems such as dim light and turbid water. With limited land area resources, the ocean must have great potential for future human development. In this paper, starting from the actual needs of marine engineering applications, taking underwater structural health monitoring (SHM) and underwater biological detection as examples, we propose improved methods for the application of underwater YOLO algorithms, and point out the problems that still exist.
Authors: Jiankai Li, Yunhong Wang, Weixin Li
Scene Graph Generation (SGG) plays a pivotal role in downstream vision-language tasks. Existing SGG methods typically suffer from poor compositional generalizations on unseen triplets. They are generally trained on incompletely annotated scene graphs that contain dominant triplets and tend to bias toward these seen triplets during inference. To address this issue, we propose a Triplet Calibration and Reduction (T-CAR) framework in this paper. In our framework, a triplet calibration loss is first presented to regularize the representations of diverse triplets and to simultaneously excavate the unseen triplets in incompletely annotated training scene graphs. Moreover, the unseen space of scene graphs is usually several times larger than the seen space since it contains a huge number of unrealistic compositions. Thus, we propose an unseen space reduction loss to shift the attention of excavation to reasonable unseen compositions to facilitate the model training. Finally, we propose a contextual encoder to improve the compositional generalizations of unseen triplets by explicitly modeling the relative spatial relations between subjects and objects. Extensive experiments show that our approach achieves consistent improvements for zero-shot SGG over state-of-the-art methods. The code is available at https://github.com/jkli1998/T-CAR.
Authors: Xiaohan Cui, Long Ma, Tengyu Ma, Jinyuan Liu, Xin Fan, Risheng Liu
Object detection in low-light scenarios has attracted much attention in the past few years. A mainstream and representative scheme introduces enhancers as the pre-processing for regular detectors. However, because of the disparity in task objectives between the enhancer and detector, this paradigm cannot shine at its best ability. In this work, we try to arouse the potential of enhancer + detector. Different from existing works, we extend the illumination-based enhancers (our newly designed or existing) as a scene decomposition module, whose removed illumination is exploited as the auxiliary in the detector for extracting detection-friendly features. A semantic aggregation module is further established for integrating multi-scale scene-related semantic information in the context space. Actually, our built scheme successfully transforms the "trash" (i.e., the ignored illumination in the detector) into the "treasure" for the detector. Plenty of experiments are conducted to reveal our superiority against other state-of-the-art methods. The code will be public if it is accepted.
Authors: Jiaxi Gu, Shicong Wang, Haoyu Zhao, Tianyi Lu, Xing Zhang, Zuxuan Wu, Songcen Xu, Wei Zhang, Yu-Gang Jiang, Hang Xu
Inspired by the remarkable success of Latent Diffusion Models (LDMs) for image synthesis, we study LDM for text-to-video generation, which is a formidable challenge due to the computational and memory constraints during both model training and inference. A single LDM is usually only capable of generating a very limited number of video frames. Some existing works focus on separate prediction models for generating more video frames, which suffer from additional training cost and frame-level jittering, however. In this paper, we propose a framework called "Reuse and Diffuse" dubbed $\textit{VidRD}$ to produce more frames following the frames already generated by an LDM. Conditioned on an initial video clip with a small number of frames, additional frames are iteratively generated by reusing the original latent features and following the previous diffusion process. Besides, for the autoencoder used for translation between pixel space and latent space, we inject temporal layers into its decoder and fine-tune these layers for higher temporal consistency. We also propose a set of strategies for composing video-text data that involve diverse content from multiple existing datasets including video datasets for action recognition and image-text datasets. Extensive experiments show that our method achieves good results in both quantitative and qualitative evaluations. Our project page is available $\href{https://anonymous0x233.github.io/ReuseAndDiffuse/}{here}$.
Authors: Sungwon Hwang, Junha Hyung, Jaegul Choo
Recent advances in diffusion models such as ControlNet have enabled geometrically controllable, high-fidelity text-to-image generation. However, none of them addresses the question of adding such controllability to text-to-3D generation. In response, we propose Text2Control3D, a controllable text-to-3D avatar generation method whose facial expression is controllable given a monocular video casually captured with hand-held camera. Our main strategy is to construct the 3D avatar in Neural Radiance Fields (NeRF) optimized with a set of controlled viewpoint-aware images that we generate from ControlNet, whose condition input is the depth map extracted from the input video. When generating the viewpoint-aware images, we utilize cross-reference attention to inject well-controlled, referential facial expression and appearance via cross attention. We also conduct low-pass filtering of Gaussian latent of the diffusion model in order to ameliorate the viewpoint-agnostic texture problem we observed from our empirical analysis, where the viewpoint-aware images contain identical textures on identical pixel positions that are incomprehensible in 3D. Finally, to train NeRF with the images that are viewpoint-aware yet are not strictly consistent in geometry, our approach considers per-image geometric variation as a view of deformation from a shared 3D canonical space. Consequently, we construct the 3D avatar in a canonical space of deformable NeRF by learning a set of per-image deformation via deformation field table. We demonstrate the empirical results and discuss the effectiveness of our method.
Authors: Shuting He, Weihua Chen, Kai Wang, Hao Luo, Fan Wang, Wei Jiang, Henghui Ding
Person Re-identification (ReID) plays a more and more crucial role in recent years with a wide range of applications. Existing ReID methods are suffering from the challenges of misalignment and occlusions, which degrade the performance dramatically. Most methods tackle such challenges by utilizing external tools to locate body parts or exploiting matching strategies. Nevertheless, the inevitable domain gap between the datasets utilized for external tools and the ReID datasets and the complicated matching process make these methods unreliable and sensitive to noises. In this paper, we propose a Region Generation and Assessment Network (RGANet) to effectively and efficiently detect the human body regions and highlight the important regions. In the proposed RGANet, we first devise a Region Generation Module (RGM) which utilizes the pre-trained CLIP to locate the human body regions using semantic prototypes extracted from text descriptions. Learnable prompt is designed to eliminate domain gap between CLIP datasets and ReID datasets. Then, to measure the importance of each generated region, we introduce a Region Assessment Module (RAM) that assigns confidence scores to different regions and reduces the negative impact of the occlusion regions by lower scores. The RAM consists of a discrimination-aware indicator and an invariance-aware indicator, where the former indicates the capability to distinguish from different identities and the latter represents consistency among the images of the same class of human body regions. Extensive experimental results for six widely-used benchmarks including three tasks (occluded, partial, and holistic) demonstrate the superiority of RGANet against state-of-the-art methods.
Authors: Luping Rao, Chuan Ma, Ming Ding, Yuwen Qian, Lu Zhou, Zhe Liu
As an essential component part of the Intelligent Transportation System (ITS), the Internet of Vehicles (IoV) plays a vital role in alleviating traffic issues. Object detection is one of the key technologies in the IoV, which has been widely used to provide traffic management services by analyzing timely and sensitive vehicle-related information. However, the current object detection methods are mostly based on centralized deep training, that is, the sensitive data obtained by edge devices need to be uploaded to the server, which raises privacy concerns. To mitigate such privacy leakage, we first propose a federated learning-based framework, where well-trained local models are shared in the central server. However, since edge devices usually have limited computing power, plus a strict requirement of low latency in IoVs, we further propose a sparse training process on edge devices, which can effectively lighten the model, and ensure its training efficiency on edge devices, thereby reducing communication overheads. In addition, due to the diverse computing capabilities and dynamic environment, different sparsity rates are applied to edge devices. To further guarantee the performance, we propose, FedWeg, an improved aggregation scheme based on FedAvg, which is designed by the inverse ratio of sparsity rates. Experiments on the real-life dataset using YOLO show that the proposed scheme can achieve the required object detection rate while saving considerable communication costs.
Authors: Yue Wang, Jinlong Peng, Jiangning Zhang, Ran Yi, Liang Liu, Yabiao Wang, Chengjie Wang
Face analysis tasks have a wide range of applications, but the universal facial representation has only been explored in a few works. In this paper, we explore high-performance pre-training methods to boost the face analysis tasks such as face alignment and face parsing. We propose a self-supervised pre-training framework, called \textbf{\it Mask Contrastive Face (MCF)}, with mask image modeling and a contrastive strategy specially adjusted for face domain tasks. To improve the facial representation quality, we use feature map of a pre-trained visual backbone as a supervision item and use a partially pre-trained decoder for mask image modeling. To handle the face identity during the pre-training stage, we further use random masks to build contrastive learning pairs. We conduct the pre-training on the LAION-FACE-cropped dataset, a variants of LAION-FACE 20M, which contains more than 20 million face images from Internet websites. For efficiency pre-training, we explore our framework pre-training performance on a small part of LAION-FACE-cropped and verify the superiority with different pre-training settings. Our model pre-trained with the full pre-training dataset outperforms the state-of-the-art methods on multiple downstream tasks. Our model achieves 0.932 NME$_{diag}$ for AFLW-19 face alignment and 93.96 F1 score for LaPa face parsing. Code is available at https://github.com/nomewang/MCF.
Authors: Haochen Wang, Junsong Fan, Yuxi Wang, Kaiyou Song, Tong Wang, Zhaoxiang Zhang
As it is empirically observed that Vision Transformers (ViTs) are quite insensitive to the order of input tokens, the need for an appropriate self-supervised pretext task that enhances the location awareness of ViTs is becoming evident. To address this, we present DropPos, a novel pretext task designed to reconstruct Dropped Positions. The formulation of DropPos is simple: we first drop a large random subset of positional embeddings and then the model classifies the actual position for each non-overlapping patch among all possible positions solely based on their visual appearance. To avoid trivial solutions, we increase the difficulty of this task by keeping only a subset of patches visible. Additionally, considering there may be different patches with similar visual appearances, we propose position smoothing and attentive reconstruction strategies to relax this classification problem, since it is not necessary to reconstruct their exact positions in these cases. Empirical evaluations of DropPos show strong capabilities. DropPos outperforms supervised pre-training and achieves competitive results compared with state-of-the-art self-supervised alternatives on a wide range of downstream benchmarks. This suggests that explicitly encouraging spatial reasoning abilities, as DropPos does, indeed contributes to the improved location awareness of ViTs. The code is publicly available at https://github.com/Haochen-Wang409/DropPos.
Authors: Vamshi K. Kancharala, Debanjali Bhattacharya, Neelam Sinha
Functional MRI (fMRI) is widely used to examine brain functionality by detecting alteration in oxygenated blood flow that arises with brain activity. In this study, complexity specific image categorization across different visual datasets is performed using fMRI time series (TS) to understand differences in neuronal activities related to vision. Publicly available BOLD5000 dataset is used for this purpose, containing fMRI scans while viewing 5254 images of diverse categories, drawn from three standard computer vision datasets: COCO, ImageNet and SUN. To understand vision, it is important to study how brain functions while looking at different images. To achieve this, spatial encoding of fMRI BOLD TS has been performed that uses classical Gramian Angular Field (GAF) and Markov Transition Field (MTF) to obtain 2D BOLD TS, representing images of COCO, Imagenet and SUN. For classification, individual GAF and MTF features are fed into regular CNN. Subsequently, parallel CNN model is employed that uses combined 2D features for classifying images across COCO, Imagenet and SUN. The result of 2D CNN models is also compared with 1D LSTM and Bi-LSTM that utilizes raw fMRI BOLD signal for classification. It is seen that parallel CNN model outperforms other network models with an improvement of 7% for multi-class classification. Clinical relevance- The obtained result of this analysis establishes a baseline in studying how differently human brain functions while looking at images of diverse complexities.
Authors: Guan Gui, Zhen Zhao, Lei Qi, Luping Zhou, Lei Wang, Yinghuan Shi
In semi-supervised learning, unlabeled samples can be utilized through augmentation and consistency regularization. However, we observed certain samples, even undergoing strong augmentation, are still correctly classified with high confidence, resulting in a loss close to zero. It indicates that these samples have been already learned well and do not provide any additional optimization benefits to the model. We refer to these samples as ``naive samples". Unfortunately, existing SSL models overlook the characteristics of naive samples, and they just apply the same learning strategy to all samples. To further optimize the SSL model, we emphasize the importance of giving attention to naive samples and augmenting them in a more diverse manner. Sample adaptive augmentation (SAA) is proposed for this stated purpose and consists of two modules: 1) sample selection module; 2) sample augmentation module. Specifically, the sample selection module picks out {naive samples} based on historical training information at each epoch, then the naive samples will be augmented in a more diverse manner in the sample augmentation module. Thanks to the extreme ease of implementation of the above modules, SAA is advantageous for being simple and lightweight. We add SAA on top of FixMatch and FlexMatch respectively, and experiments demonstrate SAA can significantly improve the models. For example, SAA helped improve the accuracy of FixMatch from 92.50% to 94.76% and that of FlexMatch from 95.01% to 95.31% on CIFAR-10 with 40 labels.
Authors: Yichen Ouyang, Wenhao Chai, Jiayi Ye, Dapeng Tao, Yibing Zhan, Gaoang Wang
Text-to-3D generation from a single-view image is a popular but challenging task in 3D vision. Although numerous methods have been proposed, existing works still suffer from the inconsistency issues, including 1) semantic inconsistency, 2) geometric inconsistency, and 3) saturation inconsistency, resulting in distorted, overfitted, and over-saturated generations. In light of the above issues, we present Consist3D, a three-stage framework Chasing for semantic-, geometric-, and saturation-Consistent Text-to-3D generation from a single image, in which the first two stages aim to learn parameterized consistency tokens, and the last stage is for optimization. Specifically, the semantic encoding stage learns a token independent of views and estimations, promoting semantic consistency and robustness. Meanwhile, the geometric encoding stage learns another token with comprehensive geometry and reconstruction constraints under novel-view estimations, reducing overfitting and encouraging geometric consistency. Finally, the optimization stage benefits from the semantic and geometric tokens, allowing a low classifier-free guidance scale and therefore preventing oversaturation. Experimental results demonstrate that Consist3D produces more consistent, faithful, and photo-realistic 3D assets compared to previous state-of-the-art methods. Furthermore, Consist3D also allows background and object editing through text prompts.
Authors: Marcello Davide Caio (1), Gabriel Van Zandycke (1 and 2), Christophe De Vleeschouwer (2) ((1) Sportradar AG, (2) UCLouvain)
Accurately localizing objects in three dimensions (3D) is crucial for various computer vision applications, such as robotics, autonomous driving, and augmented reality. This task finds another important application in sports analytics and, in this work, we present a novel method for 3D basketball localization from a single calibrated image. Our approach predicts the object's height in pixels in image space by estimating its projection onto the ground plane within the image, leveraging the image itself and the object's location as inputs. The 3D coordinates of the ball are then reconstructed by exploiting the known projection matrix. Extensive experiments on the public DeepSport dataset, which provides ground truth annotations for 3D ball location alongside camera calibration information for each image, demonstrate the effectiveness of our method, offering substantial accuracy improvements compared to recent work. Our work opens up new possibilities for enhanced ball tracking and understanding, advancing computer vision in diverse domains. The source code of this work is made publicly available at \url{https://github.com/gabriel-vanzandycke/deepsport}.
Authors: Yu Du, Xu Liu, Yansong Chua
Speech enhancement seeks to extract clean speech from noisy signals. Traditional deep learning methods face two challenges: efficiently using information in long speech sequences and high computational costs. To address these, we introduce the Spiking Structured State Space Model (Spiking-S4). This approach merges the energy efficiency of Spiking Neural Networks (SNN) with the long-range sequence modeling capabilities of Structured State Space Models (S4), offering a compelling solution. Evaluation on the DNS Challenge and VoiceBank+Demand Datasets confirms that Spiking-S4 rivals existing Artificial Neural Network (ANN) methods but with fewer computational resources, as evidenced by reduced parameters and Floating Point Operations (FLOPs).
Authors: Balint Kovacs, Nils Netzer, Michael Baumgartner, Carolin Eith, Dimitrios Bounias, Clara Meinzer, Paul F. Jaeger, Kevin S. Zhang, Ralf Floca, Adrian Schrader, Fabian Isensee, Regula Gnirs, Magdalena Goertz, Viktoria Schuetz, Albrecht Stenzinger, Markus Hohenfellner, Heinz-Peter Schlemmer, Ivo Wolf, David Bonekamp, Klaus H. Maier-Hein
Data augmentation (DA) is a key factor in medical image analysis, such as in prostate cancer (PCa) detection on magnetic resonance images. State-of-the-art computer-aided diagnosis systems still rely on simplistic spatial transformations to preserve the pathological label post transformation. However, such augmentations do not substantially increase the organ as well as tumor shape variability in the training set, limiting the model's ability to generalize to unseen cases with more diverse localized soft-tissue deformations. We propose a new anatomy-informed transformation that leverages information from adjacent organs to simulate typical physiological deformations of the prostate and generates unique lesion shapes without altering their label. Due to its lightweight computational requirements, it can be easily integrated into common DA frameworks. We demonstrate the effectiveness of our augmentation on a dataset of 774 biopsy-confirmed examinations, by evaluating a state-of-the-art method for PCa detection with different augmentation settings.
Authors: Onno Niemann, Christopher Vox, Thorben Werner
Knowledge Distillation (KD) is one proposed solution to large model sizes and slow inference speed in semantic segmentation. In our research we identify 25 proposed distillation loss terms from 14 publications in the last 4 years. Unfortunately, a comparison of terms based on published results is often impossible, because of differences in training configurations. A good illustration of this problem is the comparison of two publications from 2022. Using the same models and dataset, Structural and Statistical Texture Distillation (SSTKD) reports an increase of student mIoU of 4.54 and a final performance of 29.19, while Adaptive Perspective Distillation (APD) only improves student performance by 2.06 percentage points, but achieves a final performance of 39.25. The reason for such extreme differences is often a suboptimal choice of hyperparameters and a resulting underperformance of the student model used as reference point. In our work, we reveal problems of insufficient hyperparameter tuning by showing that distillation improvements of two widely accepted frameworks, SKD and IFVD, vanish when hyperparameters are optimized sufficiently. To improve comparability of future research in the field, we establish a solid baseline for three datasets and two student models and provide extensive information on hyperparameter tuning. We find that only two out of eight techniques can compete with our simple baseline on the ADE20K dataset.
Authors: Ting Liu, Wansen Wu, Yue Hu, Youkai Wang, Kai Xu, Quanjun Yin
With strong representation capabilities, pretrained vision-language models are widely used in vision and language navigation (VLN). However, most of them are trained on web-crawled general-purpose datasets, which incurs a considerable domain gap when used for VLN tasks. Another challenge for VLN is how the agent understands the contextual relations between actions on a trajectory and performs cross-modal alignment sequentially. In this paper, we propose a novel Prompt-bAsed coNtext- and Domain-Aware (PANDA) pretraining framework to address these problems. It performs prompting in two stages. In the domain-aware stage, we apply a low-cost prompt tuning paradigm to learn soft visual prompts from an in-domain dataset for equipping the pretrained models with object-level and scene-level cross-modal alignment in VLN tasks. Furthermore, in the context-aware stage, we design a set of hard context prompts to capture the sequence-level semantics and instill both out-of-context and contextual knowledge in the instruction into cross-modal representations. They enable further tuning of the pretrained models via contrastive learning. Experimental results on both R2R and REVERIE show the superiority of PANDA compared to previous state-of-the-art methods.
Authors: Pierre-Etienne Martin
This paper presents a bonobo detection and classification pipeline built from the commonly used machine learning methods. Such application is motivated by the need to test bonobos in their enclosure using touch screen devices without human assistance. This work introduces a newly acquired dataset based on bonobo recordings generated semi-automatically. The recordings are weakly labelled and fed to a macaque detector in order to spatially detect the individual present in the video. Handcrafted features coupled with different classification algorithms and deep-learning methods using a ResNet architecture are investigated for bonobo identification. Performance is compared in terms of classification accuracy on the splits of the database using different data separation methods. We demonstrate the importance of data preparation and how a wrong data separation can lead to false good results. Finally, after a meaningful separation of the data, the best classification performance is obtained using a fine-tuned ResNet model and reaches 75% of accuracy.
Authors: Haoyuan Chen, Yufei Han, Pin Xu, Yanyi Li, Kuan Li, Jianping Yin
Models based on U-like structures have improved the performance of medical image segmentation. However, the single-layer decoder structure of U-Net is too "thin" to exploit enough information, resulting in large semantic differences between the encoder and decoder parts. Things get worse if the number of training sets of data is not sufficiently large, which is common in medical image processing tasks where annotated data are more difficult to obtain than other tasks. Based on this observation, we propose a novel U-Net model named MS-UNet for the medical image segmentation task in this study. Instead of the single-layer U-Net decoder structure used in Swin-UNet and TransUnet, we specifically design a multi-scale nested decoder based on the Swin Transformer for U-Net. The proposed multi-scale nested decoder structure allows the feature mapping between the decoder and encoder to be semantically closer, thus enabling the network to learn more detailed features. In addition, we propose a novel edge loss and a plug-and-play fine-tuning Denoising module, which not only effectively improves the segmentation performance of MS-UNet, but could also be applied to other models individually. Experimental results show that MS-UNet could effectively improve the network performance with more efficient feature learning capability and exhibit more advanced performance, especially in the extreme case with a small amount of training data, and the proposed Edge loss and Denoising module could significantly enhance the segmentation performance of MS-UNet.
Authors: Ting Lei, Fabian Caba, Qingchao Chen, Hailin Jin, Yuxin Peng, Yang Liu
Human Object Interaction (HOI) detection aims to localize and infer the relationships between a human and an object. Arguably, training supervised models for this task from scratch presents challenges due to the performance drop over rare classes and the high computational cost and time required to handle long-tailed distributions of HOIs in complex HOI scenes in realistic settings. This observation motivates us to design an HOI detector that can be trained even with long-tailed labeled data and can leverage existing knowledge from pre-trained models. Inspired by the powerful generalization ability of the large Vision-Language Models (VLM) on classification and retrieval tasks, we propose an efficient Adaptive HOI Detector with Concept-guided Memory (ADA-CM). ADA-CM has two operating modes. The first mode makes it tunable without learning new parameters in a training-free paradigm. Its second mode incorporates an instance-aware adapter mechanism that can further efficiently boost performance if updating a lightweight set of parameters can be afforded. Our proposed method achieves competitive results with state-of-the-art on the HICO-DET and V-COCO datasets with much less training time. Code can be found at https://github.com/ltttpku/ADA-CM.
Authors: Hondamunige Prasanna Silva, Lorenzo Seidenari, Alberto Del Bimbo
This paper presents a novel reconstruction method that leverages Diffusion Models to protect machine learning classifiers against adversarial attacks, all without requiring any modifications to the classifiers themselves. The susceptibility of machine learning models to minor input perturbations renders them vulnerable to adversarial attacks. While diffusion-based methods are typically disregarded for adversarial defense due to their slow reverse process, this paper demonstrates that our proposed method offers robustness against adversarial threats while preserving clean accuracy, speed, and plug-and-play compatibility. Code at: https://github.com/HondamunigePrasannaSilva/DiffDefence.
Authors: Li Li, Qingqing Li, Guozheng Xu, Pengwei Zhou, Jingmin Tu, Jie Li, Jian Yao
Roof plane segmentation from airborne LiDAR point clouds is an important technology for 3D building model reconstruction. One of the key issues of plane segmentation is how to design powerful features that can exactly distinguish adjacent planar patches. The quality of point feature directly determines the accuracy of roof plane segmentation. Most of existing approaches use handcrafted features to extract roof planes. However, the abilities of these features are relatively low, especially in boundary area. To solve this problem, we propose a boundary-aware point clustering approach in Euclidean and embedding spaces constructed by a multi-task deep network for roof plane segmentation. We design a three-branch network to predict semantic labels, point offsets and extract deep embedding features. In the first branch, we classify the input data as non-roof, boundary and plane points. In the second branch, we predict point offsets for shifting each point toward its respective instance center. In the third branch, we constrain that points of the same plane instance should have the similar embeddings. We aim to ensure that points of the same plane instance are close as much as possible in both Euclidean and embedding spaces. However, although deep network has strong feature representative ability, it is still hard to accurately distinguish points near plane instance boundary. Therefore, we first group plane points into many clusters in the two spaces, and then we assign the rest boundary points to their closest clusters to generate final complete roof planes. In this way, we can effectively reduce the influence of unreliable boundary points. In addition, we construct a synthetic dataset and a real dataset to train and evaluate our approach. The experiments results show that the proposed approach significantly outperforms the existing state-of-the-art approaches.
Authors: Maria Parelli, Dimitrios Mallis, Markos Diomataris, Vassilis Pitsikalis
Transformer-based architectures have recently demonstrated remarkable performance in the Visual Question Answering (VQA) task. However, such models are likely to disregard crucial visual cues and often rely on multimodal shortcuts and inherent biases of the language modality to predict the correct answer, a phenomenon commonly referred to as lack of visual grounding. In this work, we alleviate this shortcoming through a novel architecture for visual question answering that leverages common sense reasoning as a supervisory signal. Reasoning supervision takes the form of a textual justification of the correct answer, with such annotations being already available on large-scale Visual Common Sense Reasoning (VCR) datasets. The model's visual attention is guided toward important elements of the scene through a similarity loss that aligns the learned attention distributions guided by the question and the correct reasoning. We demonstrate both quantitatively and qualitatively that the proposed approach can boost the model's visual perception capability and lead to performance increase, without requiring training on explicit grounding annotations.
Authors: Teng Hu, Jiangning Zhang, Liang Liu, Ran Yi, Siqi Kou, Haokun Zhu, Xu Chen, Yabiao Wang, Chengjie Wang, Lizhuang Ma
Training a generative model with limited number of samples is a challenging task. Current methods primarily rely on few-shot model adaption to train the network. However, in scenarios where data is extremely limited (less than 10), the generative network tends to overfit and suffers from content degradation. To address these problems, we propose a novel phasic content fusing few-shot diffusion model with directional distribution consistency loss, which targets different learning objectives at distinct training stages of the diffusion model. Specifically, we design a phasic training strategy with phasic content fusion to help our model learn content and style information when t is large, and learn local details of target domain when t is small, leading to an improvement in the capture of content, style and local details. Furthermore, we introduce a novel directional distribution consistency loss that ensures the consistency between the generated and source distributions more efficiently and stably than the prior methods, preventing our model from overfitting. Finally, we propose a cross-domain structure guidance strategy that enhances structure consistency during domain adaptation. Theoretical analysis, qualitative and quantitative experiments demonstrate the superiority of our approach in few-shot generative model adaption tasks compared to state-of-the-art methods. The source code is available at: https://github.com/sjtuplayer/few-shot-diffusion.
Authors: Irfan Tito Kurniawan, Bambang Riyanto Trilaksono
Thanks to the complementary nature of millimeter wave radar and camera, deep learning-based radar-camera 3D object detection methods may reliably produce accurate detections even in low-visibility conditions. This makes them preferable to use in autonomous vehicles' perception systems, especially as the combined cost of both sensors is cheaper than the cost of a lidar. Recent radar-camera methods commonly perform feature-level fusion which often involves projecting the radar points onto the same plane as the image features and fusing the extracted features from both modalities. While performing fusion on the image plane is generally simpler and faster, projecting radar points onto the image plane flattens the depth dimension of the point cloud which might lead to information loss and makes extracting the spatial features of the point cloud harder. We proposed ClusterFusion, an architecture that leverages the local spatial features of the radar point cloud by clustering the point cloud and performing feature extraction directly on the point cloud clusters before projecting the features onto the image plane. ClusterFusion achieved the state-of-the-art performance among all radar-monocular camera methods on the test slice of the nuScenes dataset with 48.7% nuScenes detection score (NDS). We also investigated the performance of different radar feature extraction strategies on point cloud clusters: a handcrafted strategy, a learning-based strategy, and a combination of both, and found that the handcrafted strategy yielded the best performance. The main goal of this work is to explore the use of radar's local spatial and point-wise features by extracting them directly from radar point cloud clusters for a radar-monocular camera 3D object detection method that performs cross-modal feature fusion on the image plane.
Authors: Nazanin Moradinasab, Rebecca A. Deaton, Laura S. Shankman, Gary K. Owens, Donald E. Brown
Recently, deep learning-based methods achieved promising performance in nuclei detection and classification applications. However, training deep learning-based methods requires a large amount of pixel-wise annotated data, which is time-consuming and labor-intensive, especially in 3D images. An alternative approach is to adapt weak-annotation methods, such as labeling each nucleus with a point, but this method does not extend from 2D histopathology images (for which it was originally developed) to 3D immunofluorescent images. The reason is that 3D images contain multiple channels (z-axis) for nuclei and different markers separately, which makes training using point annotations difficult. To address this challenge, we propose the Label-efficient Contrastive learning-based (LECL) model to detect and classify various types of nuclei in 3D immunofluorescent images. Previous methods use Maximum Intensity Projection (MIP) to convert immunofluorescent images with multiple slices to 2D images, which can cause signals from different z-stacks to falsely appear associated with each other. To overcome this, we devised an Extended Maximum Intensity Projection (EMIP) approach that addresses issues using MIP. Furthermore, we performed a Supervised Contrastive Learning (SCL) approach for weakly supervised settings. We conducted experiments on cardiovascular datasets and found that our proposed framework is effective and efficient in detecting and classifying various types of nuclei in 3D immunofluorescent images.
Authors: Sepideh Afshar, Nachiket Deo, Akshay Bhagat, Titas Chakraborty, Yunming Shao, Balarama Raju Buddharaju, Adwait Deshpande, Henggang Cui
Trajectory prediction plays a crucial role in the autonomous driving stack by enabling autonomous vehicles to anticipate the motion of surrounding agents. Goal-based prediction models have gained traction in recent years for addressing the multimodal nature of future trajectories. Goal-based prediction models simplify multimodal prediction by first predicting 2D goal locations of agents and then predicting trajectories conditioned on each goal. However, a single 2D goal location serves as a weak inductive bias for predicting the whole trajectory, often leading to poor map compliance, i.e., part of the trajectory going off-road or breaking traffic rules. In this paper, we improve upon goal-based prediction by proposing the Path-based prediction (PBP) approach. PBP predicts a discrete probability distribution over reference paths in the HD map using the path features and predicts trajectories in the path-relative Frenet frame. We applied the PBP trajectory decoder on top of the HiVT scene encoder and report results on the Argoverse dataset. Our experiments show that PBP achieves competitive performance on the standard trajectory prediction metrics, while significantly outperforming state-of-the-art baselines in terms of map compliance.
Authors: Ece Ozkan, Thomas M. Sutter, Yurong Hu, Sebastian Balzer, Julia E. Vogt
Early detection of cardiac dysfunction through routine screening is vital for diagnosing cardiovascular diseases. An important metric of cardiac function is the left ventricular ejection fraction (EF), where lower EF is associated with cardiomyopathy. Echocardiography is a popular diagnostic tool in cardiology, with ultrasound being a low-cost, real-time, and non-ionizing technology. However, human assessment of echocardiograms for calculating EF is time-consuming and expertise-demanding, raising the need for an automated approach. In this work, we propose using the M(otion)-mode of echocardiograms for estimating the EF and classifying cardiomyopathy. We generate multiple artificial M-mode images from a single echocardiogram and combine them using off-the-shelf model architectures. Additionally, we extend contrastive learning (CL) to cardiac imaging to learn meaningful representations from exploiting structures in unlabeled data allowing the model to achieve high accuracy, even with limited annotations. Our experiments show that the supervised setting converges with only ten modes and is comparable to the baseline method while bypassing its cumbersome training process and being computationally much more efficient. Furthermore, CL using M-mode images is helpful for limited data scenarios, such as having labels for only 200 patients, which is common in medical applications.
Authors: Johannes Flotzinger, Philipp J. Rösch, Norbert Oswald, Thomas Braml
Recognising reinforced concrete defects (RCDs) is a crucial element for determining the structural integrity, traffic safety and durability of bridges. However, most of the existing datasets in the RCD domain are derived from a small number of bridges acquired in specific camera poses, lighting conditions and with fixed hardware. These limitations question the usability of models trained on such open-source data in real-world scenarios. We address this problem by testing such models on our "dacl1k" dataset, a highly diverse RCD dataset for multi-label classification based on building inspections including 1,474 images. Thereby, we trained the models on different combinations of open-source data (meta datasets) which were subsequently evaluated both extrinsically and intrinsically. During extrinsic evaluation, we report metrics on dacl1k and the meta datasets. The performance analysis on dacl1k shows practical usability of the meta data, where the best model shows an Exact Match Ratio of 32%. Additionally, we conduct an intrinsic evaluation by clustering the bottleneck features of the best model derived from the extrinsic evaluation in order to find out, if the model has learned distinguishing datasets or the classes (RCDs) which is the aspired goal. The dacl1k dataset and our trained models will be made publicly available, enabling researchers and practitioners to put their models to the real-world test.
Authors: Juan Han, Kit Ian Kou, Jifei Miao, Lizhi Liu, Haojiang Li
Color image completion is a challenging problem in computer vision, but recent research has shown that quaternion representations of color images perform well in many areas. These representations consider the entire color image and effectively utilize coupling information between the three color channels. Consequently, low-rank quaternion matrix completion (LRQMC) algorithms have gained significant attention. We propose a method based on quaternion Qatar Riyal decomposition (QQR) and quaternion $L_{2,1}$-norm called QLNM-QQR. This new approach reduces computational complexity by avoiding the need to calculate the QSVD of large quaternion matrices. We also present two improvements to the QLNM-QQR method: an enhanced version called IRQLNM-QQR that uses iteratively reweighted quaternion $L_{2,1}$-norm minimization and a method called QLNM-QQR-SR that integrates sparse regularization. Our experiments on natural color images and color medical images show that IRQLNM-QQR outperforms QLNM-QQR and that the proposed QLNM-QQR-SR method is superior to several state-of-the-art methods.
Authors: Stephanie Abrecht, Alexander Hirsch, Shervin Raafatnia, Matthias Woehrle
Recent advances in the field of deep learning and impressive performance of deep neural networks (DNNs) for perception have resulted in an increased demand for their use in automated driving (AD) systems. The safety of such systems is of utmost importance and thus requires to consider the unique properties of DNNs.
In order to achieve safety of AD systems with DNN-based perception components in a systematic and comprehensive approach, so-called safety concerns have been introduced as a suitable structuring element. On the one hand, the concept of safety concerns is -- by design -- well aligned to existing standards relevant for safety of AD systems such as ISO 21448 (SOTIF). On the other hand, it has already inspired several academic publications and upcoming standards on AI safety such as ISO PAS 8800.
While the concept of safety concerns has been previously introduced, this paper extends and refines it, leveraging feedback from various domain and safety experts in the field. In particular, this paper introduces an additional categorization for a better understanding as well as enabling cross-functional teams to jointly address the concerns.
Authors: Linh Trinh, Bach Ha, Tu Tran
In many parts of the world, the use of vast amounts of data collected on public roadways for autonomous driving has increased. In order to detect and anonymize pedestrian faces and nearby car license plates in actual road-driving scenarios, there is an urgent need for effective solutions. As more data is collected, privacy concerns regarding it increase, including but not limited to pedestrian faces and surrounding vehicle license plates. Normal and fisheye cameras are the two common camera types that are typically mounted on collection vehicles. With complex camera distortion models, fisheye camera images were deformed in contrast to regular images. It causes computer vision tasks to perform poorly when using numerous deep learning models. In this work, we pay particular attention to protecting privacy while yet adhering to several laws for fisheye camera photos taken by driverless vehicles. First, we suggest a framework for extracting face and plate identification knowledge from several teacher models. Our second suggestion is to transform both the image and the label from a regular image to fisheye-like data using a varied and realistic fisheye transformation. Finally, we run a test using the open-source PP4AV dataset. The experimental findings demonstrated that our model outperformed baseline methods when trained on data from autonomous vehicles, even when the data were softly labeled. The implementation code is available at our github: https://github.com/khaclinh/FisheyePP4AV.
Authors: Christopher Wewer, Eddy Ilg, Bernt Schiele, Jan Eric Lenssen
Existing neural field representations for 3D object reconstruction either (1) utilize object-level representations, but suffer from low-quality details due to conditioning on a global latent code, or (2) are able to perfectly reconstruct the observations, but fail to utilize object-level prior knowledge to infer unobserved regions. We present SimNP, a method to learn category-level self-similarities, which combines the advantages of both worlds by connecting neural point radiance fields with a category-level self-similarity representation. Our contribution is two-fold. (1) We design the first neural point representation on a category level by utilizing the concept of coherent point clouds. The resulting neural point radiance fields store a high level of detail for locally supported object regions. (2) We learn how information is shared between neural points in an unconstrained and unsupervised fashion, which allows to derive unobserved regions of an object during the reconstruction process from given observations. We show that SimNP is able to outperform previous methods in reconstructing symmetric unseen object regions, surpassing methods that build upon category-level or pixel-aligned radiance fields, while providing semantic correspondences between instances
Authors: Sacha Jungerman, Atul Ingle, Mohit Gupta
Scene reconstruction in the presence of high-speed motion and low illumination is important in many applications such as augmented and virtual reality, drone navigation, and autonomous robotics. Traditional motion estimation techniques fail in such conditions, suffering from too much blur in the presence of high-speed motion and strong noise in low-light conditions. Single-photon cameras have recently emerged as a promising technology capable of capturing hundreds of thousands of photon frames per second thanks to their high speed and extreme sensitivity. Unfortunately, traditional computer vision techniques are not well suited for dealing with the binary-valued photon data captured by these cameras because these are corrupted by extreme Poisson noise. Here we present a method capable of estimating extreme scene motion under challenging conditions, such as low light or high dynamic range, from a sequence of high-speed image frames such as those captured by a single-photon camera. Our method relies on iteratively improving a motion estimate by grouping and aggregating frames after-the-fact, in a stratified manner. We demonstrate the creation of high-quality panoramas under fast motion and extremely low light, and super-resolution results using a custom single-photon camera prototype. For code and supplemental material see our $\href{https://wisionlab.com/project/panoramas-from-photons/}{\text{project webpage}}$.
Authors: Francesco Picetti, Shrinath Deshpande, Jonathan Leban, Soroosh Shahtalebi, Jay Patel, Peifeng Jing, Chunpu Wang, Charles Metze III, Cameron Sun, Cera Laidlaw, James Warren, Kathy Huynh, River Page, Jonathan Hogins, Adam Crespi, Sujoy Ganguly, Salehe Erfanian Ebadi
We present a novel human body model formulated by an extensive set of anthropocentric measurements, which is capable of generating a wide range of human body shapes and poses. The proposed model enables direct modeling of specific human identities through a deep generative architecture, which can produce humans in any arbitrary pose. It is the first of its kind to have been trained end-to-end using only synthetically generated data, which not only provides highly accurate human mesh representations but also allows for precise anthropometry of the body. Moreover, using a highly diverse animation library, we articulated our synthetic humans' body and hands to maximize the diversity of the learnable priors for model training. Our model was trained on a dataset of $100k$ procedurally-generated posed human meshes and their corresponding anthropometric measurements. Our synthetic data generator can be used to generate millions of unique human identities and poses for non-commercial academic research purposes.
Authors: An-An Liu, Guokai Zhang, Yuting Su, Ning Xu, Yongdong Zhang, Lanjun Wang
Recent developments in text-conditioned image generative models have revolutionized the production of realistic results. Unfortunately, this has also led to an increase in privacy violations and the spread of false information, which requires the need for traceability, privacy protection, and other security measures. However, existing text-to-image paradigms lack the technical capabilities to link traceable messages with image generation. In this study, we introduce a novel task for the joint generation of text to image and watermark (T2IW). This T2IW scheme ensures minimal damage to image quality when generating a compound image by forcing the semantic feature and the watermark signal to be compatible in pixels. Additionally, by utilizing principles from Shannon information theory and non-cooperative game theory, we are able to separate the revealed image and the revealed watermark from the compound image. Furthermore, we strengthen the watermark robustness of our approach by subjecting the compound image to various post-processing attacks, with minimal pixel distortion observed in the revealed watermark. Extensive experiments have demonstrated remarkable achievements in image quality, watermark invisibility, and watermark robustness, supported by our proposed set of evaluation metrics.
Authors: Hrishav Bakul Barua, Ganesh Krishnasamy, KokSheik Wong, Kalin Stefanov, Abhinav Dhall
High Dynamic Range (HDR) content creation has become an important topic for modern media and entertainment sectors, gaming and Augmented/Virtual Reality industries. Many methods have been proposed to recreate the HDR counterparts of input Low Dynamic Range (LDR) images/videos given a single exposure or multi-exposure LDRs. The state-of-the-art methods focus primarily on the preservation of the reconstruction's structural similarity and the pixel-wise accuracy. However, these conventional approaches do not emphasize preserving the artistic intent of the images in terms of human visual perception, which is an essential element in media, entertainment and gaming. In this paper, we attempt to study and fill this gap. We propose an architecture called ArtHDR-Net based on a Convolutional Neural Network that uses multi-exposed LDR features as input. Experimental results show that ArtHDR-Net can achieve state-of-the-art performance in terms of the HDR-VDP-2 score (i.e., mean opinion score index) while reaching competitive performance in terms of PSNR and SSIM.
Authors: Sangwook Kim, Thomas G. Purdie, Chris McIntosh
Multi-task learning (MTL) is a powerful approach in deep learning that leverages the information from multiple tasks during training to improve model performance. In medical imaging, MTL has shown great potential to solve various tasks. However, existing MTL architectures in medical imaging are limited in sharing information across tasks, reducing the potential performance improvements of MTL. In this study, we introduce a novel attention-based MTL framework to better leverage inter-task interactions for various tasks from pixel-level to image-level predictions. Specifically, we propose a Cross-Task Attention Network (CTAN) which utilizes cross-task attention mechanisms to incorporate information by interacting across tasks. We validated CTAN on four medical imaging datasets that span different domains and tasks including: radiation treatment planning prediction using planning CT images of two different target cancers (Prostate, OpenKBP); pigmented skin lesion segmentation and diagnosis using dermatoscopic images (HAM10000); and COVID-19 diagnosis and severity prediction using chest CT scans (STOIC). Our study demonstrates the effectiveness of CTAN in improving the accuracy of medical imaging tasks. Compared to standard single-task learning (STL), CTAN demonstrated a 4.67% improvement in performance and outperformed both widely used MTL baselines: hard parameter sharing (HPS) with an average performance improvement of 3.22%; and multi-task attention network (MTAN) with a relative decrease of 5.38%. These findings highlight the significance of our proposed MTL framework in solving medical imaging tasks and its potential to improve their accuracy across domains.
Authors: Ahmed H. Shahin, An Zhao, Alexander C. Whitehead, Daniel C. Alexander, Joseph Jacob, David Barber
Survival analysis is a valuable tool for estimating the time until specific events, such as death or cancer recurrence, based on baseline observations. This is particularly useful in healthcare to prognostically predict clinically important events based on patient data. However, existing approaches often have limitations; some focus only on ranking patients by survivability, neglecting to estimate the actual event time, while others treat the problem as a classification task, ignoring the inherent time-ordered structure of the events. Furthermore, the effective utilization of censored samples - training data points where the exact event time is unknown - is essential for improving the predictive accuracy of the model. In this paper, we introduce CenTime, a novel approach to survival analysis that directly estimates the time to event. Our method features an innovative event-conditional censoring mechanism that performs robustly even when uncensored data is scarce. We demonstrate that our approach forms a consistent estimator for the event model parameters, even in the absence of uncensored data. Furthermore, CenTime is easily integrated with deep learning models with no restrictions on batch size or the number of uncensored samples. We compare our approach with standard survival analysis methods, including the Cox proportional-hazard model and DeepHit. Our results indicate that CenTime offers state-of-the-art performance in predicting time-to-death while maintaining comparable ranking performance. Our implementation is publicly available at https://github.com/ahmedhshahin/CenTime.
Authors: Otniel-Bogdan Mercea, Thomas Hummel, A. Sophia Koepke, Zeynep Akata
Training deep learning models for video classification from audio-visual data commonly requires immense amounts of labeled training data collected via a costly process. A challenging and underexplored, yet much cheaper, setup is few-shot learning from video data. In particular, the inherently multi-modal nature of video data with sound and visual information has not been leveraged extensively for the few-shot video classification task. Therefore, we introduce a unified audio-visual few-shot video classification benchmark on three datasets, i.e. the VGGSound-FSL, UCF-FSL, ActivityNet-FSL datasets, where we adapt and compare ten methods. In addition, we propose AV-DIFF, a text-to-feature diffusion framework, which first fuses the temporal and audio-visual features via cross-modal attention and then generates multi-modal features for the novel classes. We show that AV-DIFF obtains state-of-the-art performance on our proposed benchmark for audio-visual (generalised) few-shot learning. Our benchmark paves the way for effective audio-visual classification when only limited labeled data is available. Code and data are available at https://github.com/ExplainableML/AVDIFF-GFSL.
Authors: Eyal Gomel, Tal Shaharabany, Lior Wolf
It has been established that training a box-based detector network can enhance the localization performance of weakly supervised and unsupervised methods. Moreover, we extend this understanding by demonstrating that these detectors can be utilized to improve the original network, paving the way for further advancements. To accomplish this, we train the detectors on top of the network output instead of the image data and apply suitable loss backpropagation. Our findings reveal a significant improvement in phrase grounding for the ``what is where by looking'' task, as well as various methods of unsupervised object discovery. Our code is available at https://github.com/eyalgomel/box-based-refinement.
Authors: Linus Ericsson, Da Li, Timothy M. Hospedales
Distribution shifts are all too common in real-world applications of machine learning. Domain adaptation (DA) aims to address this by providing various frameworks for adapting models to the deployment data without using labels. However, the domain shift scenario raises a second more subtle challenge: the difficulty of performing hyperparameter optimisation (HPO) for these adaptation algorithms without access to a labelled validation set. The unclear validation protocol for DA has led to bad practices in the literature, such as performing HPO using the target test labels when, in real-world scenarios, they are not available. This has resulted in over-optimism about DA research progress compared to reality. In this paper, we analyse the state of DA when using good evaluation practice, by benchmarking a suite of candidate validation criteria and using them to assess popular adaptation algorithms. We show that there are challenges across all three branches of domain adaptation methodology including Unsupervised Domain Adaptation (UDA), Source-Free Domain Adaptation (SFDA), and Test Time Adaptation (TTA). While the results show that realistically achievable performance is often worse than expected, they also show that using proper validation splits is beneficial, as well as showing that some previously unexplored validation metrics provide the best options to date. Altogether, our improved practices covering data, training, validation and hyperparameter optimisation form a new rigorous pipeline to improve benchmarking, and hence research progress, within this important field going forward.
Authors: Hui Zhang, Sammy Christen, Zicong Fan, Luocheng Zheng, Jemin Hwangbo, Jie Song, Otmar Hilliges
We present ArtiGrasp, a novel method to synthesize bi-manual hand-object interactions that include grasping and articulation. This task is challenging due to the diversity of the global wrist motions and the precise finger control that are necessary to articulate objects. ArtiGrasp leverages reinforcement learning and physics simulations to train a policy that controls the global and local hand pose. Our framework unifies grasping and articulation within a single policy guided by a single hand pose reference. Moreover, to facilitate the training of the precise finger control required for articulation, we present a learning curriculum with increasing difficulty. It starts with single-hand manipulation of stationary objects and continues with multi-agent training including both hands and non-stationary objects. To evaluate our method, we introduce Dynamic Object Grasping and Articulation, a task that involves bringing an object into a target articulated pose. This task requires grasping, relocation, and articulation. We show our method's efficacy towards this task. We further demonstrate that our method can generate motions with noisy hand-object pose estimates from an off-the-shelf image-based regressor.
Authors: Manlin Zhang, Jie Wu, Yuxi Ren, Ming Li, Jie Qin, Xuefeng Xiao, Wei Liu, Rui Wang, Min Zheng, Andy J. Ma
Data is the cornerstone of deep learning. This paper reveals that the recently developed Diffusion Model is a scalable data engine for object detection. Existing methods for scaling up detection-oriented data often require manual collection or generative models to obtain target images, followed by data augmentation and labeling to produce training pairs, which are costly, complex, or lacking diversity. To address these issues, we presentDiffusionEngine (DE), a data scaling-up engine that provides high-quality detection-oriented training pairs in a single stage. DE consists of a pre-trained diffusion model and an effective Detection-Adapter, contributing to generating scalable, diverse and generalizable detection data in a plug-and-play manner. Detection-Adapter is learned to align the implicit semantic and location knowledge in off-the-shelf diffusion models with detection-aware signals to make better bounding-box predictions. Additionally, we contribute two datasets, i.e., COCO-DE and VOC-DE, to scale up existing detection benchmarks for facilitating follow-up research. Extensive experiments demonstrate that data scaling-up via DE can achieve significant improvements in diverse scenarios, such as various detection algorithms, self-supervised pre-training, data-sparse, label-scarce, cross-domain, and semi-supervised learning. For example, when using DE with a DINO-based adapter to scale up data, mAP is improved by 3.1% on COCO, 7.6% on VOC, and 11.5% on Clipart.
Authors: Zigang Geng, Binxin Yang, Tiankai Hang, Chen Li, Shuyang Gu, Ting Zhang, Jianmin Bao, Zheng Zhang, Han Hu, Dong Chen, Baining Guo
We present InstructDiffusion, a unifying and generic framework for aligning computer vision tasks with human instructions. Unlike existing approaches that integrate prior knowledge and pre-define the output space (e.g., categories and coordinates) for each vision task, we cast diverse vision tasks into a human-intuitive image-manipulating process whose output space is a flexible and interactive pixel space. Concretely, the model is built upon the diffusion process and is trained to predict pixels according to user instructions, such as encircling the man's left shoulder in red or applying a blue mask to the left car. InstructDiffusion could handle a variety of vision tasks, including understanding tasks (such as segmentation and keypoint detection) and generative tasks (such as editing and enhancement). It even exhibits the ability to handle unseen tasks and outperforms prior methods on novel datasets. This represents a significant step towards a generalist modeling interface for vision tasks, advancing artificial general intelligence in the field of computer vision.
Authors: Shangchen Zhou, Chongyi Li, Kelvin C.K. Chan, Chen Change Loy
Flow-based propagation and spatiotemporal Transformer are two mainstream mechanisms in video inpainting (VI). Despite the effectiveness of these components, they still suffer from some limitations that affect their performance. Previous propagation-based approaches are performed separately either in the image or feature domain. Global image propagation isolated from learning may cause spatial misalignment due to inaccurate optical flow. Moreover, memory or computational constraints limit the temporal range of feature propagation and video Transformer, preventing exploration of correspondence information from distant frames. To address these issues, we propose an improved framework, called ProPainter, which involves enhanced ProPagation and an efficient Transformer. Specifically, we introduce dual-domain propagation that combines the advantages of image and feature warping, exploiting global correspondences reliably. We also propose a mask-guided sparse video Transformer, which achieves high efficiency by discarding unnecessary and redundant tokens. With these components, ProPainter outperforms prior arts by a large margin of 1.46 dB in PSNR while maintaining appealing efficiency.
Authors: Hala Lamdouar, Weidi Xie, Andrew Zisserman
Not all camouflages are equally effective, as even a partially visible contour or a slight color difference can make the animal stand out and break its camouflage. In this paper, we address the question of what makes a camouflage successful, by proposing three scores for automatically assessing its effectiveness. In particular, we show that camouflage can be measured by the similarity between background and foreground features and boundary visibility. We use these camouflage scores to assess and compare all available camouflage datasets. We also incorporate the proposed camouflage score into a generative model as an auxiliary loss and show that effective camouflage images or videos can be synthesised in a scalable manner. The generated synthetic dataset is used to train a transformer-based model for segmenting camouflaged animals in videos. Experimentally, we demonstrate state-of-the-art camouflage breaking performance on the public MoCA-Mask benchmark.
Authors: Su-Kai Chen, Hung-Lin Yen, Yu-Lun Liu, Min-Hung Chen, Hou-Ning Hu, Wen-Hsiao Peng, Yen-Yu Lin
Deep learning is commonly used to reconstruct HDR images from LDR images. LDR stack-based methods are used for single-image HDR reconstruction, generating an HDR image from a deep learning-generated LDR stack. However, current methods generate the stack with predetermined exposure values (EVs), which may limit the quality of HDR reconstruction. To address this, we propose the continuous exposure value representation (CEVR), which uses an implicit function to generate LDR images with arbitrary EVs, including those unseen during training. Our approach generates a continuous stack with more images containing diverse EVs, significantly improving HDR reconstruction. We use a cycle training strategy to supervise the model in generating continuous EV LDR images without corresponding ground truths. Our CEVR model outperforms existing methods, as demonstrated by experimental results.
Authors: Ho Kei Cheng, Seoung Wug Oh, Brian Price, Alexander Schwing, Joon-Young Lee
Training data for video segmentation are expensive to annotate. This impedes extensions of end-to-end algorithms to new video segmentation tasks, especially in large-vocabulary settings. To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation. Due to this design, we only need an image-level model for the target task (which is cheaper to train) and a universal temporal propagation model which is trained once and generalizes across tasks. To effectively combine these two modules, we use bi-directional propagation for (semi-)online fusion of segmentation hypotheses from different frames to generate a coherent segmentation. We show that this decoupled formulation compares favorably to end-to-end approaches in several data-scarce tasks including large-vocabulary video panoptic segmentation, open-world video segmentation, referring video segmentation, and unsupervised video object segmentation. Code is available at: https://hkchengrex.github.io/Tracking-Anything-with-DEVA
Authors: Jiapeng Zhu, Ceyuan Yang, Kecheng Zheng, Yinghao Xu, Zifan Shi, Yujun Shen
Due to the difficulty in scaling up, generative adversarial networks (GANs) seem to be falling from grace on the task of text-conditioned image synthesis. Sparsely-activated mixture-of-experts (MoE) has recently been demonstrated as a valid solution to training large-scale models with limited computational resources. Inspired by such a philosophy, we present Aurora, a GAN-based text-to-image generator that employs a collection of experts to learn feature processing, together with a sparse router to help select the most suitable expert for each feature point. To faithfully decode the sampling stochasticity and the text condition to the final synthesis, our router adaptively makes its decision by taking into account the text-integrated global latent code. At 64x64 image resolution, our model trained on LAION2B-en and COYO-700M achieves 6.2 zero-shot FID on MS COCO. We release the code and checkpoints to facilitate the community for further development.
Authors: Jiaming Han, Renrui Zhang, Wenqi Shao, Peng Gao, Peng Xu, Han Xiao, Kaipeng Zhang, Chris Liu, Song Wen, Ziyu Guo, Xudong Lu, Shuai Ren, Yafei Wen, Xiaoxin Chen, Xiangyu Yue, Hongsheng Li, Yu Qiao
We present ImageBind-LLM, a multi-modality instruction tuning method of large language models (LLMs) via ImageBind. Existing works mainly focus on language and image instruction tuning, different from which, our ImageBind-LLM can respond to multi-modality conditions, including audio, 3D point clouds, video, and their embedding-space arithmetic by only image-text alignment training. During training, we adopt a learnable bind network to align the embedding space between LLaMA and ImageBind's image encoder. Then, the image features transformed by the bind network are added to word tokens of all layers in LLaMA, which progressively injects visual instructions via an attention-free and zero-initialized gating mechanism. Aided by the joint embedding of ImageBind, the simple image-text training enables our model to exhibit superior multi-modality instruction-following capabilities. During inference, the multi-modality inputs are fed into the corresponding ImageBind encoders, and processed by a proposed visual cache model for further cross-modal embedding enhancement. The training-free cache model retrieves from three million image features extracted by ImageBind, which effectively mitigates the training-inference modality discrepancy. Notably, with our approach, ImageBind-LLM can respond to instructions of diverse modalities and demonstrate significant language generation quality. Code is released at https://github.com/OpenGVLab/LLaMA-Adapter.
Authors: Taylor W. Webb, Zachary Dulberg, Steven M. Frankland, Alexander A. Petrov, Randall C. O'Reilly, Jonathan D. Cohen
Extrapolation -- the ability to make inferences that go beyond the scope of one's experiences -- is a hallmark of human intelligence. By contrast, the generalization exhibited by contemporary neural network algorithms is largely limited to interpolation between data points in their training corpora. In this paper, we consider the challenge of learning representations that support extrapolation. We introduce a novel visual analogy benchmark that allows the graded evaluation of extrapolation as a function of distance from the convex domain defined by the training data. We also introduce a simple technique, temporal context normalization, that encourages representations that emphasize the relations between objects. We find that this technique enables a significant improvement in the ability to extrapolate, considerably outperforming a number of competitive techniques.
Authors: Jin Liang, Yuchen Yang, Anran Zhang, Jun Xu, Hui Li, Xiantong Zhen
The photographs captured by digital cameras usually suffer from over or under exposure problems. For image exposure enhancement, the tasks of Single-Exposure Correction (SEC) and Multi-Exposure Fusion (MEF) are widely studied in the image processing community. However, current SEC or MEF methods are developed under different motivations and thus ignore the internal correlation between SEC and MEF, making it difficult to process arbitrary-length sequences with improper exposures. Besides, the MEF methods usually fail at estimating the exposure of a sequence containing only under-exposed or over-exposed images. To alleviate these problems, in this paper, we develop a novel Fusion-Correction Network (FCNet) to tackle an arbitrary-length (including one) image sequence with improper exposures. This is achieved by fusing and correcting an image sequence by Laplacian Pyramid (LP) image decomposition. In each LP level, the low-frequency base component of the input image sequence is fed into a Fusion block and a Correction block sequentially for consecutive exposure estimation, implemented by alternative exposure fusion and correction. The exposure-corrected image in current LP level is upsampled and fused with the high-frequency detail components of the input image sequence in the next LP level, to output the base component for the Fusion and Correction blocks in next LP level. Experiments on the benchmark dataset demonstrate that our FCNet is effective on arbitrary-length exposure estimation, including both SEC and MEF. The code is publicly released at https://github.com/NKUJinLiang/FCNet.
Authors: Hanxing Chi, Baihong Lin, Jun Hu, Liang Wang
Recently, attention mechanisms have been extensively investigated in computer vision, but few of them show excellent performance on both large and mobile networks. This paper proposes Dual Rank-1 Tensor Attention Module (DRTAM), a novel residual-attention-learning-guided attention module for feed-forward convolutional neural networks. Given a 3D feature tensor map, DRTAM firstly generates three 2D feature descriptors along three axes. Then, using three descriptors, DRTAM sequentially infers two rank-1 tensor attention maps, the initial attention map and the complement attention map, combines and multiplied them to the input feature map for adaptive feature refinement(see Fig.1(c)). To generate two attention maps, DRTAM introduces rank-1 tensor attention module (RTAM) and residual descriptors extraction module (RDEM): RTAM divides each 2D feature descriptors into several chunks, and generate three factor vectors of a rank-1 tensor attention map by employing strip pooling on each chunk so that local and long-range contextual information can be captured along three dimension respectively; RDEM generates three 2D feature descriptors of the residual feature to produce the complement attention map, using three factor vectors of the initial attention map and three descriptors of the input feature. Extensive experimental results on ImageNet-1K, MS COCO and PASCAL VOC demonstrate that DRTAM achieves competitive performance on both large and mobile networks compare with other state-of-the-art attention modules.
Authors: Takayuki Hara, Tatsuya Harada
In this study, we present a method for synthesizing novel views from a single 360-degree RGB-D image based on the neural radiance field (NeRF) . Prior studies relied on the neighborhood interpolation capability of multi-layer perceptrons to complete missing regions caused by occlusion and zooming, which leads to artifacts. In the method proposed in this study, the input image is reprojected to 360-degree RGB images at other camera positions, the missing regions of the reprojected images are completed by a 2D image generative model, and the completed images are utilized to train the NeRF. Because multiple completed images contain inconsistencies in 3D, we introduce a method to learn the NeRF model using a subset of completed images that cover the target scene with less overlap of completed regions. The selection of such a subset of images can be attributed to the maximum weight independent set problem, which is solved through simulated annealing. Experiments demonstrated that the proposed method can synthesize plausible novel views while preserving the features of the scene for both artificial and real-world data.
Authors: Ping Zhou, Langqing Shi, Xiaoyang Liu, Jing Jin, Yuting Zhang, Junhui Hou
Depth estimation is a fundamental problem in light field processing. Epipolar-plane image (EPI)-based methods often encounter challenges such as low accuracy in slope computation due to discretization errors and limited angular resolution. Besides, existing methods perform well in most regions but struggle to produce sharp edges in occluded regions and resolve ambiguities in texture-less regions. To address these issues, we propose the concept of stitched-EPI (SEPI) to enhance slope computation. SEPI achieves this by shifting and concatenating lines from different EPIs that correspond to the same 3D point. Moreover, we introduce the half-SEPI algorithm, which focuses exclusively on the non-occluded portion of lines to handle occlusion. Additionally, we present a depth propagation strategy aimed at improving depth estimation in texture-less regions. This strategy involves determining the depth of such regions by progressing from the edges towards the interior, prioritizing accurate regions over coarse regions. Through extensive experimental evaluations and ablation studies, we validate the effectiveness of our proposed method. The results demonstrate its superior ability to generate more accurate and robust depth maps across all regions compared to state-of-the-art methods. The source code will be publicly available at https://github.com/PingZhou-LF/Light-Field-Depth-Estimation-Based-on-Stitched-EPIs.
Authors: Gang Xu (1), Yu-chen Yang (1), Liang Wang (2), Xian-Tong Zhen (3), Jun Xu (1) ((1) Nankai University, (2) Institute of Automation, CAS, (3) Guangdong University of Petrochemical Technology)
Joint Super-Resolution and Inverse Tone-Mapping (joint SR-ITM) aims to increase the resolution and dynamic range of low-resolution and standard dynamic range images. Recent networks mainly resort to image decomposition techniques with complex multi-branch architectures. However, the fixed decomposition techniques would largely restricts their power on versatile images. To exploit the potential power of decomposition mechanism, in this paper, we generalize it from the image domain to the broader feature domain. To this end, we propose a lightweight Feature Decomposition Aggregation Network (FDAN). In particular, we design a Feature Decomposition Block (FDB) to achieve learnable separation of detail and base feature maps, and develop a Hierarchical Feature Decomposition Group by cascading FDBs for powerful multi-level feature decomposition. Moreover, to better evaluate the comparison methods, we collect a large-scale dataset for joint SR-ITM, i.e., SRITM-4K, which provides versatile scenarios for robust model training and evaluation. Experimental results on two benchmark datasets demonstrate that our FDAN is efficient and outperforms state-of-the-art methods on joint SR-ITM. The code of our FDAN and the SRITM-4K dataset are available at https://github.com/CS-GangXu/FDAN.
Authors: Vinay Ummadi, Aravind Gundlapalle, Althaf Shaik, Shaik Mohammad Rafi B
This project aims to develop and demonstrate a ground robot with intelligence capable of conducting semi-autonomous farm operations for different low-heights vegetable crops referred as Agriculture Application Robot(AAR). AAR is a lightweight, solar-electric powered robot that uses intelligent perception for conducting detection and classification of plants and their characteristics. The system also has a robotic arm for the autonomous weed cutting process. The robot can deliver fertilizer spraying, insecticide, herbicide, and other fluids to the targets such as crops, weeds, and other pests. Besides, it provides information for future research into higher-level tasks such as yield estimation, crop, and soil health monitoring. We present the design of robot and the associated experiments which show the promising results in real world environments.
Authors: Sophie Ostmeier, Brian Axelrod, Jeroen Bertels, Fabian Isensee, Maarten G.Lansberg, Soren Christensen, Gregory W. Albers, Li-Jia Li, Jeremy J. Heit
Performance metrics for medical image segmentation models are used to measure the agreement between the reference annotation and the predicted segmentation. Usually, overlap metrics, such as the Dice, are used as a metric to evaluate the performance of these models in order for results to be comparable. However, there is a mismatch between the distributions of cases and difficulty level of segmentation tasks in public data sets compared to clinical practice. Common metrics fail to measure the impact of this mismatch, especially for clinical data sets that include low signal pathologies, a difficult segmentation task, and uncertain, small, or empty reference annotations. This limitation may result in ineffective research of machine learning practitioners in designing and optimizing models. Dimensions of evaluating clinical value include consideration of the uncertainty of reference annotations, independence from reference annotation volume size, and evaluation of classification of empty reference annotations. We study how uncertain, small, and empty reference annotations influence the value of metrics for medical image segmentation on an in-house data set regardless of the model. We examine metrics behavior on the predictions of a standard deep learning framework in order to identify metrics with clinical value. We compare to a public benchmark data set (BraTS 2019) with a high-signal pathology and certain, larger, and no empty reference annotations. We may show machine learning practitioners, how uncertain, small, or empty reference annotations require a rethinking of the evaluation and optimizing procedures. The evaluation code was released to encourage further analysis of this topic. https://github.com/SophieOstmeier/UncertainSmallEmpty.git
Authors: Alexandre Tiard, Alex Wong, David Joon Ho, Yangchao Wu, Eliram Nof, Alvin C. Goh, Stefano Soatto, Saad Nadeem
We present a self-supervised algorithm for several classification tasks within hematoxylin and eosin (H&E) stained images of breast cancer. Our method is robust to stain variations inherent to the histology images acquisition process, which has limited the applicability of automated analysis tools. We address this problem by imposing constraints a learnt latent space which leverages stain normalization techniques during training. At every iteration, we select an image as a normalization target and generate a version of every image in the batch normalized to that target. We minimize the distance between the embeddings that correspond to the same image under different staining variations while maximizing the distance between other samples. We show that our method not only improves robustness to stain variations across multi-center data, but also classification performance through extensive experiments on various normalization targets and methods. Our method achieves the state-of-the-art performance on several publicly available breast cancer datasets ranging from tumor classification (CAMELYON17) and subtyping (BRACS) to HER2 status classification and treatment response prediction.
Authors: Sophie Ostmeier, Brian Axelrod, Benjamin F.J. Verhaaren, Soren Christensen, Abdelkader Mahammedi, Yongkai Liu, Benjamin Pulli, Li-Jia Li, Greg Zaharchuk, Jeremy J. Heit
To determine if a convolutional neural network (CNN) deep learning model can accurately segment acute ischemic changes on non-contrast CT compared to neuroradiologists. Non-contrast CT (NCCT) examinations from 232 acute ischemic stroke patients who were enrolled in the DEFUSE 3 trial were included in this study. Three experienced neuroradiologists independently segmented hypodensity that reflected the ischemic core on each scan. The neuroradiologist with the most experience (expert A) served as the ground truth for deep learning model training. Two additional neuroradiologists (experts B and C) segmentations were used for data testing. The 232 studies were randomly split into training and test sets. The training set was further randomly divided into 5 folds with training and validation sets. A 3-dimensional CNN architecture was trained and optimized to predict the segmentations of expert A from NCCT. The performance of the model was assessed using a set of volume, overlap, and distance metrics using non-inferiority thresholds of 20%, 3ml, and 3mm. The optimized model trained on expert A was compared to test experts B and C. We used a one-sided Wilcoxon signed-rank test to test for the non-inferiority of the model-expert compared to the inter-expert agreement. The final model performance for the ischemic core segmentation task reached a performance of 0.46+-0.09 Surface Dice at Tolerance 5mm and 0.47+-0.13 Dice when trained on expert A. Compared to the two test neuroradiologists the model-expert agreement was non-inferior to the inter-expert agreement, p < 0.05. The CNN accurately delineates the hypodense ischemic core on NCCT in acute ischemic stroke patients with an accuracy comparable to neuroradiologists.
Authors: Jun Luo, Matias Mendieta, Chen Chen, Shandong Wu
Personalized federated learning has received an upsurge of attention due to the mediocre performance of conventional federated learning (FL) over heterogeneous data. Unlike conventional FL which trains a single global consensus model, personalized FL allows different models for different clients. However, existing personalized FL algorithms only implicitly transfer the collaborative knowledge across the federation by embedding the knowledge into the aggregated model or regularization. We observed that this implicit knowledge transfer fails to maximize the potential of each client's empirical risk toward other clients. Based on our observation, in this work, we propose Personalized Global Federated Learning (PGFed), a novel personalized FL framework that enables each client to personalize its own global objective by explicitly and adaptively aggregating the empirical risks of itself and other clients. To avoid massive (O(N^2)) communication overhead and potential privacy leakage while achieving this, each client's risk is estimated through a first-order approximation for other clients' adaptive risk aggregation. On top of PGFed, we develop a momentum upgrade, dubbed PGFedMo, to more efficiently utilize clients' empirical risks. Our extensive experiments on four datasets under different federated settings show consistent improvements of PGFed over previous state-of-the-art methods. The code is publicly available at https://github.com/ljaiverson/pgfed.
Authors: Guangyao Zhou, Nishad Gothoskar, Lirui Wang, Joshua B. Tenenbaum, Dan Gutfreund, Miguel Lázaro-Gredilla, Dileep George, Vikash K. Mansinghka
The ability to perceive and understand 3D scenes is crucial for many applications in computer vision and robotics. Inverse graphics is an appealing approach to 3D scene understanding that aims to infer the 3D scene structure from 2D images. In this paper, we introduce probabilistic modeling to the inverse graphics framework to quantify uncertainty and achieve robustness in 6D pose estimation tasks. Specifically, we propose 3D Neural Embedding Likelihood (3DNEL) as a unified probabilistic model over RGB-D images, and develop efficient inference procedures on 3D scene descriptions. 3DNEL effectively combines learned neural embeddings from RGB with depth information to improve robustness in sim-to-real 6D object pose estimation from RGB-D images. Performance on the YCB-Video dataset is on par with state-of-the-art yet is much more robust in challenging regimes. In contrast to discriminative approaches, 3DNEL's probabilistic generative formulation jointly models multiple objects in a scene, quantifies uncertainty in a principled way, and handles object pose tracking under heavy occlusion. Finally, 3DNEL provides a principled framework for incorporating prior knowledge about the scene and objects, which allows natural extension to additional tasks like camera pose tracking from video.
Authors: Dovile Juodelyte, Amelia Jiménez-Sánchez, Veronika Cheplygina
While a key component to the success of deep learning is the availability of massive amounts of training data, medical image datasets are often limited in diversity and size. Transfer learning has the potential to bridge the gap between related yet different domains. For medical applications, however, it remains unclear whether it is more beneficial to pre-train on natural or medical images. We aim to shed light on this problem by comparing initialization on ImageNet and RadImageNet on seven medical classification tasks. Our work includes a replication study, which yields results contrary to previously published findings. In our experiments, ResNet50 models pre-trained on ImageNet tend to outperform those trained on RadImageNet. To gain further insights, we investigate the learned representations using Canonical Correlation Analysis (CCA) and compare the predictions of the different models. Our results indicate that, contrary to intuition, ImageNet and RadImageNet may converge to distinct intermediate representations, which appear to diverge further during fine-tuning. Despite these distinct representations, the predictions of the models remain similar. Our findings show that the similarity between networks before and after fine-tuning does not correlate with performance gains, suggesting that the advantages of transfer learning might not solely originate from the reuse of features in the early layers of a convolutional neural network.
Authors: Alexander C. Li, Ellis Brown, Alexei A. Efros, Deepak Pathak
Modern vision models typically rely on fine-tuning general-purpose models pre-trained on large, static datasets. These general-purpose models only capture the knowledge within their pre-training datasets, which are tiny, out-of-date snapshots of the Internet -- where billions of images are uploaded each day. We suggest an alternate approach: rather than hoping our static datasets transfer to our desired tasks after large-scale pre-training, we propose dynamically utilizing the Internet to quickly train a small-scale model that does extremely well on the task at hand. Our approach, called Internet Explorer, explores the web in a self-supervised manner to progressively find relevant examples that improve performance on a desired target dataset. It cycles between searching for images on the Internet with text queries, self-supervised training on downloaded images, determining which images were useful, and prioritizing what to search for next. We evaluate Internet Explorer across several datasets and show that it outperforms or matches CLIP oracle performance by using just a single GPU desktop to actively query the Internet for 30--40 hours. Results, visualizations, and videos at https://internet-explorer-ssl.github.io/
Authors: Yue Wang, Jinlong Peng, Jiangning Zhang, Ran Yi, Yabiao Wang, Chengjie Wang
2D-based Industrial Anomaly Detection has been widely discussed, however, multimodal industrial anomaly detection based on 3D point clouds and RGB images still has many untouched fields. Existing multimodal industrial anomaly detection methods directly concatenate the multimodal features, which leads to a strong disturbance between features and harms the detection performance. In this paper, we propose Multi-3D-Memory (M3DM), a novel multimodal anomaly detection method with hybrid fusion scheme: firstly, we design an unsupervised feature fusion with patch-wise contrastive learning to encourage the interaction of different modal features; secondly, we use a decision layer fusion with multiple memory banks to avoid loss of information and additional novelty classifiers to make the final decision. We further propose a point feature alignment operation to better align the point cloud and RGB features. Extensive experiments show that our multimodal industrial anomaly detection model outperforms the state-of-the-art (SOTA) methods on both detection and segmentation precision on MVTec-3D AD dataset. Code is available at https://github.com/nomewang/M3DM.
Authors: Bharath Srinivas Prabakaran, Erik Ostrowski, Muhammad Shafique
Weakly Supervised Semantic Segmentation (WSSS) relying only on image-level supervision is a promising approach to deal with the need for Segmentation networks, especially for generating a large number of pixel-wise masks in a given dataset. However, most state-of-the-art image-level WSSS techniques lack an understanding of the geometric features embedded in the images since the network cannot derive any object boundary information from just image-level labels. We define a boundary here as the line separating an object and its background, or two different objects. To address this drawback, we are proposing our novel ReFit framework, which deploys state-of-the-art class activation maps combined with various post-processing techniques in order to achieve fine-grained higher-accuracy segmentation masks. To achieve this, we investigate a state-of-the-art unsupervised segmentation network that can be used to construct a boundary map, which enables ReFit to predict object locations with sharper boundaries. By applying our method to WSSS predictions, we achieved up to 10% improvement over the current state-of-the-art WSSS methods for medical imaging. The framework is open-source, to ensure that our results are reproducible, and accessible online at https://github.com/bharathprabakaran/ReFit.
Authors: Erik Ostrowski, Muhammad Shafique
One key bottleneck of employing state-of-the-art semantic segmentation networks in the real world is the availability of training labels. Conventional semantic segmentation networks require massive pixel-wise annotated labels to reach state-of-the-art prediction quality. Hence, several works focus on semantic segmentation networks trained with only image-level annotations. However, when scrutinizing the results of state-of-the-art in more detail, we notice that they are remarkably close to each other on average prediction quality, different approaches perform better in different classes while providing low quality in others. To address this problem, we propose a novel framework, ISLE, which employs an ensemble of the "pseudo-labels" for a given set of different semantic segmentation techniques on a class-wise level. Pseudo-labels are the pixel-wise predictions of the image-level semantic segmentation frameworks used to train the final segmentation model. Our pseudo-labels seamlessly combine the strong points of multiple segmentation techniques approaches to reach superior prediction quality. We reach up to 2.4% improvement over ISLE's individual components. An exhaustive analysis was performed to demonstrate ISLE's effectiveness over state-of-the-art frameworks for image-level semantic segmentation.
Authors: Shihao Zou, Yuxuan Mu, Xinxin Zuo, Sen Wang, Li Cheng
Event camera, as an emerging biologically-inspired vision sensor for capturing motion dynamics, presents new potential for 3D human pose tracking, or video-based 3D human pose estimation. However, existing works in pose tracking either require the presence of additional gray-scale images to establish a solid starting pose, or ignore the temporal dependencies all together by collapsing segments of event streams to form static event frames. Meanwhile, although the effectiveness of Artificial Neural Networks (ANNs, a.k.a. dense deep learning) has been showcased in many event-based tasks, the use of ANNs tends to neglect the fact that compared to the dense frame-based image sequences, the occurrence of events from an event camera is spatiotemporally much sparser. Motivated by the above mentioned issues, we present in this paper a dedicated end-to-end sparse deep learning approach for event-based pose tracking: 1) to our knowledge this is the first time that 3D human pose tracking is obtained from events only, thus eliminating the need of accessing to any frame-based images as part of input; 2) our approach is based entirely upon the framework of Spiking Neural Networks (SNNs), which consists of Spike-Element-Wise (SEW) ResNet and a novel Spiking Spatiotemporal Transformer; 3) a large-scale synthetic dataset is constructed that features a broad and diverse set of annotated 3D human motions, as well as longer hours of event stream data, named SynEventHPD. Empirical experiments demonstrate that, with superior performance over the state-of-the-art (SOTA) ANNs counterparts, our approach also achieves a significant computation reduction of 80% in FLOPS. Furthermore, our proposed method also outperforms SOTA SNNs in the regression task of human pose tracking. Our implementation is available at https://github.com/JimmyZou/HumanPoseTracking_SNN and dataset will be released upon paper acceptance.
Authors: Tim Lebailly, Thomas Stegmüller, Behzad Bozorgtabar, Jean-Philippe Thiran, Tinne Tuytelaars
Most self-supervised methods for representation learning leverage a cross-view consistency objective i.e., they maximize the representation similarity of a given image's augmented views. Recent work NNCLR goes beyond the cross-view paradigm and uses positive pairs from different images obtained via nearest neighbor bootstrapping in a contrastive setting. We empirically show that as opposed to the contrastive learning setting which relies on negative samples, incorporating nearest neighbor bootstrapping in a self-distillation scheme can lead to a performance drop or even collapse. We scrutinize the reason for this unexpected behavior and provide a solution. We propose to adaptively bootstrap neighbors based on the estimated quality of the latent space. We report consistent improvements compared to the naive bootstrapping approach and the original baselines. Our approach leads to performance improvements for various self-distillation method/backbone combinations and standard downstream tasks. Our code is publicly available at https://github.com/tileb1/AdaSim.
Authors: Maximilian Nielsen, Laura Wenderoth, Thilo Sentker, René Werner
Is self-supervised deep learning (DL) for medical image analysis already a serious alternative to the de facto standard of end-to-end trained supervised DL? We tackle this question for medical image classification, with a particular focus on one of the currently most limiting factors of the field: the (non-)availability of labeled data. Based on three common medical imaging modalities (bone marrow microscopy, gastrointestinal endoscopy, dermoscopy) and publicly available data sets, we analyze the performance of self-supervised DL within the self-distillation with no labels (DINO) framework. After learning an image representation without use of image labels, conventional machine learning classifiers are applied. The classifiers are fit using a systematically varied number of labeled data (1-1000 samples per class). Exploiting the learned image representation, we achieve state-of-the-art classification performance for all three imaging modalities and data sets with only a fraction of between 1% and 10% of the available labeled data and about 100 labeled samples per class.
Authors: Ilwi Yun, Chanyong Shin, Hyunku Lee, Hyuk-Jae Lee, Chae Eun Rhee
Estimating the depths of equirectangular (i.e., 360) images (EIs) is challenging given the distorted 180 x 360 field-of-view, which is hard to be addressed via convolutional neural network (CNN). Although a transformer with global attention achieves significant improvements over CNN for EI depth estimation task, it is computationally inefficient, which raises the need for transformer with local attention. However, to apply local attention successfully for EIs, a specific strategy, which addresses distorted equirectangular geometry and limited receptive field simultaneously, is required. Prior works have only cared either of them, resulting in unsatisfactory depths occasionally. In this paper, we propose an equirectangular geometry-biased transformer termed EGformer. While limiting the computational cost and the number of network parameters, EGformer enables the extraction of the equirectangular geometry-aware local attention with a large receptive field. To achieve this, we actively utilize the equirectangular geometry as the bias for the local attention instead of struggling to reduce the distortion of EIs. As compared to the most recent EI depth estimation studies, the proposed approach yields the best depth outcomes overall with the lowest computational cost and the fewest parameters, demonstrating the effectiveness of the proposed methods.
Authors: Puntawat Ponglertnapakorn, Nontawat Tritrong, Supasorn Suwajanakorn
We present a novel approach to single-view face relighting in the wild. Handling non-diffuse effects, such as global illumination or cast shadows, has long been a challenge in face relighting. Prior work often assumes Lambertian surfaces, simplified lighting models or involves estimating 3D shape, albedo, or a shadow map. This estimation, however, is error-prone and requires many training examples with lighting ground truth to generalize well. Our work bypasses the need for accurate estimation of intrinsic components and can be trained solely on 2D images without any light stage data, multi-view images, or lighting ground truth. Our key idea is to leverage a conditional diffusion implicit model (DDIM) for decoding a disentangled light encoding along with other encodings related to 3D shape and facial identity inferred from off-the-shelf estimators. We also propose a novel conditioning technique that eases the modeling of the complex interaction between light and geometry by using a rendered shading reference to spatially modulate the DDIM. We achieve state-of-the-art performance on standard benchmark Multi-PIE and can photorealistically relight in-the-wild images. Please visit our page: https://diffusion-face-relighting.github.io
Authors: Zheren Li, Zhiming Cui, Lichi Zhang, Sheng Wang, Chenjin Lei, Xi Ouyang, Dongdong Chen, Xiangyu Zhao, Yajia Gu, Zaiyi Liu, Chunling Liu, Dinggang Shen, Jie-Zhi Cheng
The deep learning technique has been shown to be effectively addressed several image analysis tasks in the computer-aided diagnosis scheme for mammography. The training of an efficacious deep learning model requires large data with diverse styles and qualities. The diversity of data often comes from the use of various scanners of vendors. But, in practice, it is impractical to collect a sufficient amount of diverse data for training. To this end, a novel contrastive learning is developed to equip the deep learning models with better style generalization capability. Specifically, the multi-style and multi-view unsupervised self-learning scheme is carried out to seek robust feature embedding against style diversity as a pretrained model. Afterward, the pretrained network is further fine-tuned to the downstream tasks, e.g., mass detection, matching, BI-RADS rating, and breast density classification. The proposed method has been evaluated extensively and rigorously with mammograms from various vendor style domains and several public datasets. The experimental results suggest that the proposed domain generalization method can effectively improve performance of four mammographic image tasks on the data from both seen and unseen domains, and outperform many state-of-the-art (SOTA) generalization methods.
Authors: Yooshin Cho, Hanbyel Cho, Hyeong Gwon Hong, Jaesung Ahn, Dongmin Cho, JungWoo Chang, Junmo Kim
Since the beginning of world-wide COVID-19 pandemic, facial masks have been recommended to limit the spread of the disease. However, these masks hide certain facial attributes. Hence, it has become difficult for existing face recognition systems to perform identity verification on masked faces. In this context, it is necessary to develop masked Face Recognition (MFR) for contactless biometric recognition systems. Thus, in this paper, we propose Complementary Attention Learning and Multi-Focal Spatial Attention that precisely removes masked region by training complementary spatial attention to focus on two distinct regions: masked regions and backgrounds. In our method, standard spatial attention and networks focus on unmasked regions, and extract mask-invariant features while minimizing the loss of the conventional Face Recognition (FR) performance. For conventional FR, we evaluate the performance on the IJB-C, Age-DB, CALFW, and CPLFW datasets. We evaluate the MFR performance on the ICCV2021-MFR/Insightface track, and demonstrate the improved performance on the both MFR and FR datasets. Additionally, we empirically verify that spatial attention of proposed method is more precisely activated in unmasked regions.
Authors: Wan Jiang, Yunfeng Diao, He Wang, Jianxin Sun, Meng Wang, Richang Hong
Safeguarding data from unauthorized exploitation is vital for privacy and security, especially in recent rampant research in security breach such as adversarial/membership attacks. To this end, \textit{unlearnable examples} (UEs) have been recently proposed as a compelling protection, by adding imperceptible perturbation to data so that models trained on them cannot classify them accurately on original clean distribution. Unfortunately, we find UEs provide a false sense of security, because they cannot stop unauthorized users from utilizing other unprotected data to remove the protection, by turning unlearnable data into learnable again. Motivated by this observation, we formally define a new threat by introducing \textit{learnable unauthorized examples} (LEs) which are UEs with their protection removed. The core of this approach is a novel purification process that projects UEs onto the manifold of LEs. This is realized by a new joint-conditional diffusion model which denoises UEs conditioned on the pixel and perceptual similarity between UEs and LEs. Extensive experiments demonstrate that LE delivers state-of-the-art countering performance against both supervised UEs and unsupervised UEs in various scenarios, which is the first generalizable countermeasure to UEs across supervised learning and unsupervised learning. Our code is available at \url{https://github.com/jiangw-0/LE_JCDP}.
Authors: Wenjin Wang, Yunhao Li, Yixin Ou, Yin Zhang
Layout-aware pre-trained models has achieved significant progress on document image question answering. They introduce extra learnable modules into existing language models to capture layout information within document images from text bounding box coordinates obtained by OCR tools. However, extra modules necessitate pre-training on extensive document images. This prevents these methods from directly utilizing off-the-shelf instruction-tuning language foundation models, which have recently shown promising potential in zero-shot learning. Instead, in this paper, we find that instruction-tuning language models like Claude and ChatGPT can understand layout by spaces and line breaks. Based on this observation, we propose the LAyout and Task aware Instruction Prompt (LATIN-Prompt), which consists of layout-aware document content and task-aware instruction. Specifically, the former uses appropriate spaces and line breaks to recover the layout information among text segments obtained by OCR tools, and the latter ensures that generated answers adhere to formatting requirements. Moreover, we propose the LAyout and Task aware Instruction Tuning (LATIN-Tuning) to improve the performance of small instruction-tuning models like Alpaca. Experimental results show that LATIN-Prompt enables zero-shot performance of Claude and ChatGPT to be comparable to the fine-tuning performance of SOTAs on document image question answering, and LATIN-Tuning enhances the zero-shot performance of Alpaca significantly. For example, LATIN-Prompt improves the performance of Claude and ChatGPT on DocVQA by 263% and 20% respectively. LATIN-Tuning improves the performance of Alpaca on DocVQA by 87.7%. Quantitative and qualitative analyses demonstrate the effectiveness of LATIN-Prompt and LATIN-Tuning. We provide the code in supplementary and will release it to facilitate future research.
Authors: Jianeng Wang, Jonathan D. Gammell
Event-based cameras asynchronously capture individual visual changes in a scene. This makes them more robust than traditional frame-based cameras to highly dynamic motions and poor illumination. It also means that every measurement in a scene can occur at a unique time.
Handling these different measurement times is a major challenge of using event-based cameras. It is often addressed in visual odometry (VO) pipelines by approximating temporally close measurements as occurring at one common time. This grouping simplifies the estimation problem but, absent additional sensors, sacrifices the inherent temporal resolution of event-based cameras.
This paper instead presents a complete stereo VO pipeline that estimates directly with individual event-measurement times without requiring any grouping or approximation in the estimation state. It uses continuous-time trajectory estimation to maintain the temporal fidelity and asynchronous nature of event-based cameras through Gaussian process regression with a physically motivated prior. Its performance is evaluated on the MVSEC dataset, where it achieves 7.9e-3 and 5.9e-3 RMS relative error on two independent sequences, outperforming the existing publicly available event-based stereo VO pipeline by two and four times, respectively.
Authors: Shahzaib Iqbal, Tariq M. Khan, Syed S. Naqvi, Muhammad Usman, Imran Razzak
In this study, we propose LDMRes-Net, a lightweight dual-multiscale residual block-based computational neural network tailored for medical image segmentation on IoT and edge platforms. Conventional U-Net-based models face challenges in meeting the speed and efficiency demands of real-time clinical applications, such as disease monitoring, radiation therapy, and image-guided surgery. LDMRes-Net overcomes these limitations with its remarkably low number of learnable parameters (0.072M), making it highly suitable for resource-constrained devices. The model's key innovation lies in its dual multi-residual block architecture, which enables the extraction of refined features on multiple scales, enhancing overall segmentation performance. To further optimize efficiency, the number of filters is carefully selected to prevent overlap, reduce training time, and improve computational efficiency. The study includes comprehensive evaluations, focusing on segmentation of the retinal image of vessels and hard exudates crucial for the diagnosis and treatment of ophthalmology. The results demonstrate the robustness, generalizability, and high segmentation accuracy of LDMRes-Net, positioning it as an efficient tool for accurate and rapid medical image segmentation in diverse clinical applications, particularly on IoT and edge platforms. Such advances hold significant promise for improving healthcare outcomes and enabling real-time medical image analysis in resource-limited settings.
Authors: Yifei Liu, Mathias Gehrig, Nico Messikommer, Marco Cannici, Davide Scaramuzza
Vision Transformers (ViTs) have shown impressive performance in computer vision, but their high computational cost, quadratic in the number of tokens, limits their adoption in computation-constrained applications. However, this large number of tokens may not be necessary, as not all tokens are equally important. In this paper, we investigate token pruning to accelerate inference for object detection and instance segmentation, extending prior works from image classification. Through extensive experiments, we offer four insights for dense tasks: (i) tokens should not be completely pruned and discarded, but rather preserved in the feature maps for later use. (ii) reactivating previously pruned tokens can further enhance model performance. (iii) a dynamic pruning rate based on images is better than a fixed pruning rate. (iv) a lightweight, 2-layer MLP can effectively prune tokens, achieving accuracy comparable with complex gating networks with a simpler design. We evaluate the impact of these design choices on COCO dataset and present a method integrating these insights that outperforms prior art token pruning models, significantly reducing performance drop from ~1.5 mAP to ~0.3 mAP for both boxes and masks. Compared to the dense counterpart that uses all tokens, our method achieves up to 34% faster inference speed for the whole network and 46% for the backbone.
Authors: Ori Gordon, Omri Avrahami, Dani Lischinski
Editing a local region or a specific object in a 3D scene represented by a NeRF or consistently blending a new realistic object into the scene is challenging, mainly due to the implicit nature of the scene representation. We present Blended-NeRF, a robust and flexible framework for editing a specific region of interest in an existing NeRF scene, based on text prompts, along with a 3D ROI box. Our method leverages a pretrained language-image model to steer the synthesis towards a user-provided text prompt, along with a 3D MLP model initialized on an existing NeRF scene to generate the object and blend it into a specified region in the original scene. We allow local editing by localizing a 3D ROI box in the input scene, and blend the content synthesized inside the ROI with the existing scene using a novel volumetric blending technique. To obtain natural looking and view-consistent results, we leverage existing and new geometric priors and 3D augmentations for improving the visual fidelity of the final result. We test our framework both qualitatively and quantitatively on a variety of real 3D scenes and text prompts, demonstrating realistic multi-view consistent results with much flexibility and diversity compared to the baselines. Finally, we show the applicability of our framework for several 3D editing applications, including adding new objects to a scene, removing/replacing/altering existing objects, and texture conversion.
Authors: Rahul Palnitkar, Jeova Farias Sales Rocha Neto
Spectral Clustering is one of the most traditional methods to solve segmentation problems. Based on Normalized Cuts, it aims at partitioning an image using an objective function defined by a graph. Despite their mathematical attractiveness, spectral approaches are traditionally neglected by the scientific community due to their practical issues and underperformance. In this paper, we adopt a sparse graph formulation based on the inclusion of extra nodes to a simple grid graph. While the grid encodes the pixel spatial disposition, the extra nodes account for the pixel color data. Applying the original Normalized Cuts algorithm to this graph leads to a simple and scalable method for spectral image segmentation, with an interpretable solution. Our experiments also demonstrate that our proposed methodology over performs both traditional and modern unsupervised algorithms for segmentation in both real and synthetic data.
Authors: Jingyu Zhuang, Chen Wang, Lingjie Liu, Liang Lin, Guanbin Li
Neural fields have achieved impressive advancements in view synthesis and scene reconstruction. However, editing these neural fields remains challenging due to the implicit encoding of geometry and texture information. In this paper, we propose DreamEditor, a novel framework that enables users to perform controlled editing of neural fields using text prompts. By representing scenes as mesh-based neural fields, DreamEditor allows localized editing within specific regions. DreamEditor utilizes the text encoder of a pretrained text-to-Image diffusion model to automatically identify the regions to be edited based on the semantics of the text prompts. Subsequently, DreamEditor optimizes the editing region and aligns its geometry and texture with the text prompts through score distillation sampling [29]. Extensive experiments have demonstrated that DreamEditor can accurately edit neural fields of real-world scenes according to the given text prompts while ensuring consistency in irrelevant areas. DreamEditor generates highly realistic textures and geometry, significantly surpassing previous works in both quantitative and qualitative evaluations.
Authors: Didi Zhu, Zexi Li, Min Zhang, Junkun Yuan, Yunfeng Shao, Jiashuo Liu, Kun Kuang, Yinchuan Li, Chao Wu
Large-scale vision-language (V-L) models have demonstrated remarkable generalization capabilities for downstream tasks through prompt tuning. However, the mechanisms behind the learned text representations are unknown, limiting further generalization gains, especially under class imbalance scenarios. Recent advances in the neural collapse (NC) phenomenon of vision-only models suggest that the optimal representation structure is the simplex ETF, which paves the way to study representations in V-L models. In this paper, we make the first attempt to use NC for examining the representations in V-L models via prompt tuning. It is found that NC optimality of text-to-image representations shows a positive correlation with downstream generalizability, which is more severe under class imbalance settings. To improve the representations, we propose Neural-collapse-anchored Prompt Tuning (NPT), a novel method that learns prompts with text and image representations that satisfy the same simplex ETF. NPT incorporates two regularization terms: language-modality collapse and multi-modality isomorphism; and it is compatible with other prompt tuning methods. Extensive experiments show that NPT can consistently help to improve existing prompt tuning techniques across 11 datasets for both balanced and imbalanced settings.
Authors: Feng Liu, Ryan Ashbaugh, Nicholas Chimitt, Najmul Hassan, Ali Hassani, Ajay Jaiswal, Minchul Kim, Zhiyuan Mao, Christopher Perry, Zhiyuan Ren, Yiyang Su, Pegah Varghaei, Kai Wang, Xingguang Zhang, Stanley Chan, Arun Ross, Humphrey Shi, Zhangyang Wang, Anil Jain, Xiaoming Liu
Whole-body biometric recognition is an important area of research due to its vast applications in law enforcement, border security, and surveillance. This paper presents the end-to-end design, development and evaluation of FarSight, an innovative software system designed for whole-body (fusion of face, gait and body shape) biometric recognition. FarSight accepts videos from elevated platforms and drones as input and outputs a candidate list of identities from a gallery. The system is designed to address several challenges, including (i) low-quality imagery, (ii) large yaw and pitch angles, (iii) robust feature extraction to accommodate large intra-person variabilities and large inter-person similarities, and (iv) the large domain gap between training and test sets. FarSight combines the physics of imaging and deep learning models to enhance image restoration and biometric feature encoding. We test FarSight's effectiveness using the newly acquired IARPA Biometric Recognition and Identification at Altitude and Range (BRIAR) dataset. Notably, FarSight demonstrated a substantial performance increase on the BRIAR dataset, with gains of +11.82% Rank-20 identification and +11.3% TAR@1% FAR.
Authors: Debaditya Roy, Dhruv Verma, Basura Fernando
Situation Recognition is the task of generating a structured summary of what is happening in an image using an activity verb and the semantic roles played by actors and objects. In this task, the same activity verb can describe a diverse set of situations as well as the same actor or object category can play a diverse set of semantic roles depending on the situation depicted in the image. Hence a situation recognition model needs to understand the context of the image and the visual-linguistic meaning of semantic roles. Therefore, we leverage the CLIP foundational model that has learned the context of images via language descriptions. We show that deeper-and-wider multi-layer perceptron (MLP) blocks obtain noteworthy results for the situation recognition task by using CLIP image and text embedding features and it even outperforms the state-of-the-art CoFormer, a Transformer-based model, thanks to the external implicit visual-linguistic knowledge encapsulated by CLIP and the expressive power of modern MLP block designs. Motivated by this, we design a cross-attention-based Transformer using CLIP visual tokens that model the relation between textual roles and visual entities. Our cross-attention-based Transformer known as ClipSitu XTF outperforms existing state-of-the-art by a large margin of 14.1\% on semantic role labelling (value) for top-1 accuracy using imSitu dataset. {Similarly, our ClipSitu XTF obtains state-of-the-art situation localization performance.} We will make the code publicly available.
Authors: Jakob Drachmann Havtorn, Amelie Royer, Tijmen Blankevoort, Babak Ehteshami Bejnordi
The input tokens to Vision Transformers carry little semantic meaning as they are defined as regular equal-sized patches of the input image, regardless of its content. However, processing uniform background areas of an image should not necessitate as much compute as dense, cluttered areas. To address this issue, we propose a dynamic mixed-scale tokenization scheme for ViT, MSViT. Our method introduces a conditional gating mechanism that selects the optimal token scale for every image region, such that the number of tokens is dynamically determined per input. In addition, to enhance the conditional behavior of the gate during training, we introduce a novel generalization of the batch-shaping loss. We show that our gating module is able to learn meaningful semantics despite operating locally at the coarse patch-level. The proposed gating module is lightweight, agnostic to the choice of transformer backbone, and trained within a few epochs with little training overhead. Furthermore, in contrast to token pruning, MSViT does not lose information about the input, thus can be readily applied for dense tasks. We validate MSViT on the tasks of classification and segmentation where it leads to improved accuracy-complexity trade-off.
Authors: Chantal Pellegrini, Matthias Keicher, Ege Özsoy, Nassir Navab
Radiology reporting is a crucial part of the communication between radiologists and other medical professionals, but it can be time-consuming and error-prone. One approach to alleviate this is structured reporting, which saves time and enables a more accurate evaluation than free-text reports. However, there is limited research on automating structured reporting, and no public benchmark is available for evaluating and comparing different methods. To close this gap, we introduce Rad-ReStruct, a new benchmark dataset that provides fine-grained, hierarchically ordered annotations in the form of structured reports for X-Ray images. We model the structured reporting task as hierarchical visual question answering (VQA) and propose hi-VQA, a novel method that considers prior context in the form of previously asked questions and answers for populating a structured radiology report. Our experiments show that hi-VQA achieves competitive performance to the state-of-the-art on the medical VQA benchmark VQARad while performing best among methods without domain-specific vision-language pretraining and provides a strong baseline on Rad-ReStruct. Our work represents a significant step towards the automated population of structured radiology reports and provides a valuable first benchmark for future research in this area. Our dataset and code is available at https://github.com/ChantalMP/Rad-ReStruct.
Authors: Guoqiang Zhang, J. P. Lewis, W. Bastiaan Kleijn
Recently, various methods have been proposed to address the inconsistency issue of DDIM inversion to enable image editing, such as EDICT [36] and Null-text inversion [22]. However, the above methods introduce considerable computational overhead. In this paper, we propose a new technique, named \emph{bi-directional integration approximation} (BDIA), to perform exact diffusion inversion with neglible computational overhead. Suppose we would like to estimate the next diffusion state $\boldsymbol{z}_{i-1}$ at timestep $t_i$ with the historical information $(i,\boldsymbol{z}_i)$ and $(i+1,\boldsymbol{z}_{i+1})$. We first obtain the estimated Gaussian noise $\hat{\boldsymbol{\epsilon}}(\boldsymbol{z}_i,i)$, and then apply the DDIM update procedure twice for approximating the ODE integration over the next time-slot $[t_i, t_{i-1}]$ in the forward manner and the previous time-slot $[t_i, t_{t+1}]$ in the backward manner. The DDIM step for the previous time-slot is used to refine the integration approximation made earlier when computing $\boldsymbol{z}_i$. A nice property of BDIA-DDIM is that the update expression for $\boldsymbol{z}_{i-1}$ is a linear combination of $(\boldsymbol{z}_{i+1}, \boldsymbol{z}_i, \hat{\boldsymbol{\epsilon}}(\boldsymbol{z}_i,i))$. This allows for exact backward computation of $\boldsymbol{z}_{i+1}$ given $(\boldsymbol{z}_i, \boldsymbol{z}_{i-1})$, thus leading to exact diffusion inversion. It is demonstrated with experiments that (round-trip) BDIA-DDIM is particularly effective for image editing. Our experiments further show that BDIA-DDIM produces markedly better image sampling qualities than DDIM for text-to-image generation.
BDIA can also be applied to improve the performance of other ODE solvers in addition to DDIM. In our work, it is found that applying BDIA to the EDM sampling procedure produces new SOTA performance over CIFAR10.
Authors: Hu Zhang, Yanchen Li, Luziwei Leng, Kaiwei Che, Qian Liu, Qinghai Guo, Jianxing Liao, Ran Cheng
Event-based sensors, distinguished by their high temporal resolution of 1 {\mu}s and a dynamic range of 120 dB, stand out as ideal tools for deployment in fast-paced settings like vehicles and drones. Traditional object detection techniques that utilize Artificial Neural Networks (ANNs) face challenges due to the sparse and asynchronous nature of the events these sensors capture. In contrast, Spiking Neural Networks (SNNs) offer a promising alternative, providing a temporal representation that is inherently aligned with event-based data. This paper explores the unique membrane potential dynamics of SNNs and their ability to modulate sparse events. We introduce an innovative spike-triggered adaptive threshold mechanism designed for stable training. Building on these insights, we present a specialized spiking feature pyramid network (SpikeFPN) optimized for automotive event based object detection. Comprehensive evaluations demonstrate that SpikeFPN surpasses both traditional SNNs and advanced ANNs enhanced with attention mechanisms. Evidently, SpikeFPN achieves a mean Average Precision (mAP) of 0.477 on the GEN1 Automotive Detection (GAD) benchmark dataset, marking a significant increase of 9.7% over the previous best SNN. Moreover, the efficient design of SpikeFPN ensures robust performance while optimizing computational resources, attributed to its innate sparse computation capabilities.
Authors: Ziyi Wang, Xumin Yu, Yongming Rao, Jie Zhou, Jiwen Lu
With the overwhelming trend of mask image modeling led by MAE, generative pre-training has shown a remarkable potential to boost the performance of fundamental models in 2D vision. However, in 3D vision, the over-reliance on Transformer-based backbones and the unordered nature of point clouds have restricted the further development of generative pre-training. In this paper, we propose a novel 3D-to-2D generative pre-training method that is adaptable to any point cloud model. We propose to generate view images from different instructed poses via the cross-attention mechanism as the pre-training scheme. Generating view images has more precise supervision than its point cloud counterpart, thus assisting 3D backbones to have a finer comprehension of the geometrical structure and stereoscopic relations of the point cloud. Experimental results have proved the superiority of our proposed 3D-to-2D generative pre-training over previous pre-training methods. Our method is also effective in boosting the performance of architecture-oriented approaches, achieving state-of-the-art performance when fine-tuning on ScanObjectNN classification and ShapeNetPart segmentation tasks. Code is available at https://github.com/wangzy22/TAP.
Authors: Qilin Zhou, Zhengyuan Wei, Haipeng Wang, W.K. Chan
Patch robustness certification ensures no patch within a given bound on a sample can manipulate a deep learning model to predict a different label. However, existing techniques cannot certify samples that cannot meet their strict bars at the classifier or patch region levels. This paper proposes MajorCert. MajorCert firstly finds all possible label sets manipulatable by the same patch region on the same sample across the underlying classifiers, then enumerates their combinations element-wise, and finally checks whether the majority invariant of all these combinations is intact to certify samples.
Authors: Zhengyang Mao, Wei Ju, Yifang Qin, Xiao Luo, Ming Zhang
Graph classification is a crucial task in many real-world multimedia applications, where graphs can represent various multimedia data types such as images, videos, and social networks. Previous efforts have applied graph neural networks (GNNs) in balanced situations where the class distribution is balanced. However, real-world data typically exhibit long-tailed class distributions, resulting in a bias towards the head classes when using GNNs and limited generalization ability over the tail classes. Recent approaches mainly focus on re-balancing different classes during model training, which fails to explicitly introduce new knowledge and sacrifices the performance of the head classes. To address these drawbacks, we propose a novel framework called Retrieval Augmented Hybrid Network (RAHNet) to jointly learn a robust feature extractor and an unbiased classifier in a decoupled manner. In the feature extractor training stage, we develop a graph retrieval module to search for relevant graphs that directly enrich the intra-class diversity for the tail classes. Moreover, we innovatively optimize a category-centered supervised contrastive loss to obtain discriminative representations, which is more suitable for long-tailed scenarios. In the classifier fine-tuning stage, we balance the classifier weights with two weight regularization techniques, i.e., Max-norm and weight decay. Experiments on various popular benchmarks verify the superiority of the proposed method against state-of-the-art approaches.
Authors: Jianyu Wen, Chenhao Wu, Tong Zhang, Yixuan Yu, Piotr Swierczynski
In this paper, we propose a 2-stage low-light image enhancement method called Self-Reference Deep Adaptive Curve Estimation (Self-DACE). In the first stage, we present an intuitive, lightweight, fast, and unsupervised luminance enhancement algorithm. The algorithm is based on a novel low-light enhancement curve that can be used to locally boost image brightness. We also propose a new loss function with a simplified physical model designed to preserve natural images' color, structure, and fidelity. We use a vanilla CNN to map each pixel through deep Adaptive Adjustment Curves (AAC) while preserving the local image structure. Secondly, we introduce the corresponding denoising scheme to remove the latent noise in the darkness. We approximately model the noise in the dark and deploy a Denoising-Net to estimate and remove the noise after the first stage. Exhaustive qualitative and quantitative analysis shows that our method outperforms existing state-of-the-art algorithms on multiple real-world datasets.
Authors: Tao Yang, Peiran Ren, Xuansong Xie, Lei Zhang
Realistic image super-resolution (Real-ISR) aims to reproduce perceptually realistic image details from a low-quality input. The commonly used adversarial training based Real-ISR methods often introduce unnatural visual artifacts and fail to generate realistic textures for natural scene images. The recently developed generative stable diffusion models provide a potential solution to Real-ISR with pre-learned strong image priors. However, the existing methods along this line either fail to keep faithful pixel-wise image structures or resort to extra skipped connections to reproduce details, which requires additional training in image space and limits their extension to other related tasks in latent space such as image stylization. In this work, we propose a pixel-aware stable diffusion (PASD) network to achieve robust Real-ISR as well as personalized stylization. In specific, a pixel-aware cross attention module is introduced to enable diffusion models perceiving image local structures in pixel-wise level, while a degradation removal module is used to extract degradation insensitive features to guide the diffusion process together with image high level information. By simply replacing the base diffusion model with a personalized one, our method can generate diverse stylized images without the need to collect pairwise training data. PASD can be easily integrated into existing diffusion models such as Stable Diffusion. Experiments on Real-ISR and personalized stylization demonstrate the effectiveness of our proposed approach. The source code and models can be found at \url{https://github.com/yangxy/PASD}.
Authors: Yuting Xiao, Jingwei Xu, Zehao Yu, Shenghua Gao
In recent years, the neural implicit surface has emerged as a powerful representation for multi-view surface reconstruction due to its simplicity and state-of-the-art performance. However, reconstructing smooth and detailed surfaces in indoor scenes from multi-view images presents unique challenges. Indoor scenes typically contain large texture-less regions, making the photometric loss unreliable for optimizing the implicit surface. Previous work utilizes monocular geometry priors to improve the reconstruction in indoor scenes. However, monocular priors often contain substantial errors in thin structure regions due to domain gaps and the inherent inconsistencies when derived independently from different views. This paper presents \textbf{DebSDF} to address these challenges, focusing on the utilization of uncertainty in monocular priors and the bias in SDF-based volume rendering. We propose an uncertainty modeling technique that associates larger uncertainties with larger errors in the monocular priors. High-uncertainty priors are then excluded from optimization to prevent bias. This uncertainty measure also informs an importance-guided ray sampling and adaptive smoothness regularization, enhancing the learning of fine structures. We further introduce a bias-aware signed distance function to density transformation that takes into account the curvature and the angle between the view direction and the SDF normals to reconstruct fine details better. Our approach has been validated through extensive experiments on several challenging datasets, demonstrating improved qualitative and quantitative results in reconstructing thin structures in indoor scenes, thereby outperforming previous work.
Authors: Christoph Reich, Biplob Debnath, Deep Patel, Tim Prangemeier, Srimat Chakradhar
Lossy video compression is commonly used when transmitting and storing video data. Unified video codecs (e.g., H.264 or H.265) remain the de facto standard, despite the availability of advanced (neural) compression approaches. Transmitting videos in the face of dynamic network bandwidth conditions requires video codecs to adapt to vastly different compression strengths. Rate control modules augment the codec's compression such that bandwidth constraints are satisfied and video distortion is minimized. While, both standard video codes and their rate control modules are developed to minimize video distortion w.r.t. human quality assessment, preserving the downstream performance of deep vision models is not considered. In this paper, we present the first end-to-end learnable deep video codec control considering both bandwidth constraints and downstream vision performance, while not breaking existing standardization. We demonstrate for two common vision tasks (semantic segmentation and optical flow estimation) and on two different datasets that our deep codec control better preserves downstream performance than using 2-pass average bit rate control while meeting dynamic bandwidth constraints and adhering to standardizations.
Authors: Xin Li, Wenqing Chu, Ye Wu, Weihang Yuan, Fanglong Liu, Qi Zhang, Fu Li, Haocheng Feng, Errui Ding, Jingdong Wang
In this paper, we present VideoGen, a text-to-video generation approach, which can generate a high-definition video with high frame fidelity and strong temporal consistency using reference-guided latent diffusion. We leverage an off-the-shelf text-to-image generation model, e.g., Stable Diffusion, to generate an image with high content quality from the text prompt, as a reference image to guide video generation. Then, we introduce an efficient cascaded latent diffusion module conditioned on both the reference image and the text prompt, for generating latent video representations, followed by a flow-based temporal upsampling step to improve the temporal resolution. Finally, we map latent video representations into a high-definition video through an enhanced video decoder. During training, we use the first frame of a ground-truth video as the reference image for training the cascaded latent diffusion module. The main characterises of our approach include: the reference image generated by the text-to-image model improves the visual fidelity; using it as the condition makes the diffusion model focus more on learning the video dynamics; and the video decoder is trained over unlabeled video data, thus benefiting from high-quality easily-available videos. VideoGen sets a new state-of-the-art in text-to-video generation in terms of both qualitative and quantitative evaluation. See \url{https://videogen.github.io/VideoGen/} for more samples.
Authors: Sanyi Zhang, Xiaochun Cao, Rui Wang, Guo-Jun Qi, Jie Zhou
Human parsing aims to segment each pixel of the human image with fine-grained semantic categories. However, current human parsers trained with clean data are easily confused by numerous image corruptions such as blur and noise. To improve the robustness of human parsers, in this paper, we construct three corruption robustness benchmarks, termed LIP-C, ATR-C, and Pascal-Person-Part-C, to assist us in evaluating the risk tolerance of human parsing models. Inspired by the data augmentation strategy, we propose a novel heterogeneous augmentation-enhanced mechanism to bolster robustness under commonly corrupted conditions. Specifically, two types of data augmentations from different views, i.e., image-aware augmentation and model-aware image-to-image transformation, are integrated in a sequential manner for adapting to unforeseen image corruptions. The image-aware augmentation can enrich the high diversity of training images with the help of common image operations. The model-aware augmentation strategy that improves the diversity of input data by considering the model's randomness. The proposed method is model-agnostic, and it can plug and play into arbitrary state-of-the-art human parsing frameworks. The experimental results show that the proposed method demonstrates good universality which can improve the robustness of the human parsing models and even the semantic segmentation models when facing various image common corruptions. Meanwhile, it can still obtain approximate performance on clean data.
Authors: Minsu Kim, Yongjun Lee, Woo Kyoung Han, Kyong Hwan Jin
Trendy suggestions for learning-based elastic warps enable the deep image stitchings to align images exposed to large parallax errors. Despite the remarkable alignments, the methods struggle with occasional holes or discontinuity between overlapping and non-overlapping regions of a target image as the applied training strategy mostly focuses on overlap region alignment. As a result, they require additional modules such as seam finder and image inpainting for hiding discontinuity and filling holes, respectively. In this work, we suggest Recurrent Elastic Warps (REwarp) that address the problem with Dirichlet boundary condition and boost performances by residual learning for recurrent misalign correction. Specifically, REwarp predicts a homography and a Thin-plate Spline (TPS) under the boundary constraint for discontinuity and hole-free image stitching. Our experiments show the favorable aligns and the competitive computational costs of REwarp compared to the existing stitching methods. Our source code is available at https://github.com/minshu-kim/REwarp.
Authors: Minsu Kim, Jaewon Lee, Byeonghun Lee, Sunghoon Im, Kyong Hwan Jin
Existing frameworks for image stitching often provide visually reasonable stitchings. However, they suffer from blurry artifacts and disparities in illumination, depth level, etc. Although the recent learning-based stitchings relax such disparities, the required methods impose sacrifice of image qualities failing to capture high-frequency details for stitched images. To address the problem, we propose a novel approach, implicit Neural Image Stitching (NIS) that extends arbitrary-scale super-resolution. Our method estimates Fourier coefficients of images for quality-enhancing warps. Then, the suggested model blends color mismatches and misalignment in the latent space and decodes the features into RGB values of stitched images. Our experiments show that our approach achieves improvement in resolving the low-definition imaging of the previous deep image stitching with favorable accelerated image-enhancing methods. Our source code is available at https://github.com/minshu-kim/NIS.
Authors: Fengfan Zhou, Hefei Ling, Yuxuan Shi, Jiazhong Chen, Ping Li
Adversarial face examples possess two critical properties: Visual Quality and Transferability. However, existing approaches rarely address these properties simultaneously, leading to subpar results. To address this issue, we propose a novel adversarial attack technique known as Adversarial Restoration (AdvRestore), which enhances both visual quality and transferability of adversarial face examples by leveraging a face restoration prior. In our approach, we initially train a Restoration Latent Diffusion Model (RLDM) designed for face restoration. Subsequently, we employ the inference process of RLDM to generate adversarial face examples. The adversarial perturbations are applied to the intermediate features of RLDM. Additionally, by treating RLDM face restoration as a sibling task, the transferability of the generated adversarial face examples is further improved. Our experimental results validate the effectiveness of the proposed attack method.
Authors: Yuanshuo Cheng, Mingwen Shao, Yecong Wan, Lixu Zhang, Wangmeng Zuo, Deyu Meng
Existing Video Restoration (VR) methods always necessitate the individual deployment of models for each adverse weather to remove diverse adverse weather degradations, lacking the capability for adaptive processing of degradations. Such limitation amplifies the complexity and deployment costs in practical applications. To overcome this deficiency, in this paper, we propose a Cross-consistent Deep Unfolding Network (CDUN) for All-In-One VR, which enables the employment of a single model to remove diverse degradations for the first time. Specifically, the proposed CDUN accomplishes a novel iterative optimization framework, capable of restoring frames corrupted by corresponding degradations according to the degradation features given in advance. To empower the framework for eliminating diverse degradations, we devise a Sequence-wise Adaptive Degradation Estimator (SADE) to estimate degradation features for the input corrupted video. By orchestrating these two cascading procedures, CDUN achieves adaptive processing for diverse degradation. In addition, we introduce a window-based inter-frame fusion strategy to utilize information from more adjacent frames. This strategy involves the progressive stacking of temporal windows in multiple iterations, effectively enlarging the temporal receptive field and enabling each frame's restoration to leverage information from distant frames. Extensive experiments demonstrate that the proposed method achieves state-of-the-art performance in All-In-One VR.
Authors: Zhangyong Tang, Tianyang Xu, Xuefeng Zhu, Xiao-Jun Wu, Josef Kittler
Generative models (GMs) have received increasing research interest for their remarkable capacity to achieve comprehensive understanding. However, their potential application in the domain of multi-modal tracking has remained relatively unexplored. In this context, we seek to uncover the potential of harnessing generative techniques to address the critical challenge, information fusion, in multi-modal tracking. In this paper, we delve into two prominent GM techniques, namely, Conditional Generative Adversarial Networks (CGANs) and Diffusion Models (DMs). Different from the standard fusion process where the features from each modality are directly fed into the fusion block, we condition these multi-modal features with random noise in the GM framework, effectively transforming the original training samples into harder instances. This design excels at extracting discriminative clues from the features, enhancing the ultimate tracking performance. To quantitatively gauge the effectiveness of our approach, we conduct extensive experiments across two multi-modal tracking tasks, three baseline methods, and three challenging benchmarks. The experimental results demonstrate that the proposed generative-based fusion mechanism achieves state-of-the-art performance, setting new records on LasHeR and RGBD1K.
Authors: Teru Nagamori, Sayaka Shiota, Hitoshi Kiya
In recent years, deep neural networks (DNNs) trained with transformed data have been applied to various applications such as privacy-preserving learning, access control, and adversarial defenses. However, the use of transformed data decreases the performance of models. Accordingly, in this paper, we propose a novel method for fine-tuning models with transformed images under the use of the vision transformer (ViT). The proposed domain adaptation method does not cause the accuracy degradation of models, and it is carried out on the basis of the embedding structure of ViT. In experiments, we confirmed that the proposed method prevents accuracy degradation even when using encrypted images with the CIFAR-10 and CIFAR-100 datasets.
Authors: Guang Yang, Yin Tang, Zhijian Wu, Jun Li, Jianhua Xu, Xili Wan
Recent mainstream masked distillation methods function by reconstructing selectively masked areas of a student network from the feature map of its teacher counterpart. In these methods, the masked regions need to be properly selected, such that reconstructed features encode sufficient discrimination and representation capability like the teacher feature. However, previous masked distillation methods only focus on spatial masking, making the resulting masked areas biased towards spatial importance without encoding informative channel clues. In this study, we devise a Dual Masked Knowledge Distillation (DMKD) framework which can capture both spatially important and channel-wise informative clues for comprehensive masked feature reconstruction. More specifically, we employ dual attention mechanism for guiding the respective masking branches, leading to reconstructed feature encoding dual significance. Furthermore, fusing the reconstructed features is achieved by self-adjustable weighting strategy for effective feature distillation. Our experiments on object detection task demonstrate that the student networks achieve performance gains of 4.1% and 4.3% with the help of our method when RetinaNet and Cascade Mask R-CNN are respectively used as the teacher networks, while outperforming the other state-of-the-art distillation methods.
Authors: Shanzhi Yin, Tongda Xu, Yongsheng Liang, Yuanyuan Wang, Yanghao Li, Yan Wang, Jingjing Liu
With neural networks growing deeper and feature maps growing larger, limited communication bandwidth with external memory (or DRAM) and power constraints become a bottleneck in implementing network inference on mobile and edge devices. In this paper, we propose an end-to-end differentiable bandwidth efficient neural inference method with the activation compressed by neural data compression method. Specifically, we propose a transform-quantization-entropy coding pipeline for activation compression with symmetric exponential Golomb coding and a data-dependent Gaussian entropy model for arithmetic coding. Optimized with existing model quantization methods, low-level task of image compression can achieve up to 19x bandwidth reduction with 6.21x energy saving.
Authors: Ada Gorgun, Yeti Z. Gurbuz, A. Aydin Alatan
Typical technique in knowledge distillation (KD) is regularizing the learning of a limited capacity model (student) by pushing its responses to match a powerful model's (teacher). Albeit useful especially in the penultimate layer and beyond, its action on student's feature transform is rather implicit, limiting its practice in the intermediate layers. To explicitly embed the teacher's knowledge in feature transform, we propose a learnable KD layer for the student which improves KD with two distinct abilities: i) learning how to leverage the teacher's knowledge, enabling to discard nuisance information, and ii) feeding forward the transferred knowledge deeper. Thus, the student enjoys the teacher's knowledge during the inference besides training. Formally, we repurpose 1x1-BN-ReLU-1x1 convolution block to assign a semantic vector to each local region according to the template (supervised by the teacher) that the corresponding region of the student matches. To facilitate template learning in the intermediate layers, we propose a novel form of supervision based on the teacher's decisions. Through rigorous experimentation, we demonstrate the effectiveness of our approach on 3 popular classification benchmarks. Code is available at: https://github.com/adagorgun/letKD-framework