Authors: Pajon Quentin, Serre Swan, Wissocq Hugo, Rabaud Léo, Haidar Siba, Yaacoub Antoun
This paper presents an investigation into machine learning techniques for violence detection in videos and their adaptation to a federated learning context. The study includes experiments with spatio-temporal features extracted from benchmark video datasets, comparison of different methods, and proposal of a modified version of the "Flow-Gated" architecture called "Diff-Gated." Additionally, various machine learning techniques, including super-convergence and transfer learning, are explored, and a method for adapting centralized datasets to a federated learning context is developed. The research achieves better accuracy results compared to state-of-the-art models by training the best violence detection model in a federated learning context.
Authors: Ruiqi Zhang, Jie Chen, Qiang Wang
This paper proposes a technique for efficiently modeling dynamic humans by explicifying the implicit neural fields via a Neural Explicit Surface (NES). Implicit neural fields have advantages over traditional explicit representations in modeling dynamic 3D content from sparse observations and effectively representing complex geometries and appearances. Implicit neural fields defined in 3D space, however, are expensive to render due to the need for dense sampling during volumetric rendering. Moreover, their memory efficiency can be further optimized when modeling sparse 3D space. To overcome these issues, the paper proposes utilizing Neural Explicit Surface (NES) to explicitly represent implicit neural fields, facilitating memory and computational efficiency. To achieve this, the paper creates a fully differentiable conversion between the implicit neural fields and the explicit rendering interface of NES, leveraging the strengths of both implicit and explicit approaches. This conversion enables effective training of the hybrid representation using implicit methods and efficient rendering by integrating the explicit rendering interface with a newly proposed rasterization-based neural renderer that only incurs a texture color query once for the initial ray interaction with the explicit surface, resulting in improved inference efficiency. NES describes dynamic human geometries with pose-dependent neural implicit surface deformation fields and their dynamic neural textures both in 2D space, which is a more memory-efficient alternative to traditional 3D methods, reducing redundancy and computational load. The comprehensive experiments show that NES performs similarly to previous 3D approaches, with greatly improved rendering speed and reduced memory cost.
Authors: Nicha C. Dvornek, Catherine Sullivan, James S. Duncan, Abha R. Gupta
The multifactorial etiology of autism spectrum disorder (ASD) suggests that its study would benefit greatly from multimodal approaches that combine data from widely varying platforms, e.g., neuroimaging, genetics, and clinical characterization. Prior neuroimaging-genetic analyses often apply naive feature concatenation approaches in data-driven work or use the findings from one modality to guide posthoc analysis of another, missing the opportunity to analyze the paired multimodal data in a truly unified approach. In this paper, we develop a more integrative model for combining genetic, demographic, and neuroimaging data. Inspired by the influence of genotype on phenotype, we propose using an attention-based approach where the genetic data guides attention to neuroimaging features of importance for model prediction. The genetic data is derived from copy number variation parameters, while the neuroimaging data is from functional magnetic resonance imaging. We evaluate the proposed approach on ASD classification and severity prediction tasks, using a sex-balanced dataset of 228 ASD and typically developing subjects in a 10-fold cross-validation framework. We demonstrate that our attention-based model combining genetic information, demographic data, and functional magnetic resonance imaging results in superior prediction performance compared to other multimodal approaches.
Authors: Yuanhan Mo, Yao Chen, Aimee Readie, Gregory Ligozio, Thibaud Coroller, Bartłomiej W. Papież
Manually grading structural changes with the modified Stoke Ankylosing Spondylitis Spinal Score (mSASSS) on spinal X-ray imaging is costly and time-consuming due to bone shape complexity and image quality variations. In this study, we address this challenge by prototyping a 2-step auto-grading pipeline, called VertXGradeNet, to automatically predict mSASSS scores for the cervical and lumbar vertebral units (VUs) in X-ray spinal imaging. The VertXGradeNet utilizes VUs generated by our previously developed VU extraction pipeline (VertXNet) as input and predicts mSASSS based on those VUs. VertXGradeNet was evaluated on an in-house dataset of lateral cervical and lumbar X-ray images for axial spondylarthritis patients. Our results show that VertXGradeNet can predict the mSASSS score for each VU when the data is limited in quantity and imbalanced. Overall, it can achieve a balanced accuracy of 0.56 and 0.51 for 4 different mSASSS scores (i.e., a score of 0, 1, 2, 3) on two test datasets. The accuracy of the presented method shows the potential to streamline the spinal radiograph readings and therefore reduce the cost of future clinical trials.
Authors: Harshit Shah, Aravindhan G, Pavan Kulkarni, Yuvaraj Govidarajulu, Manojkumar Parmar
A significant number of machine learning models are vulnerable to model extraction attacks, which focus on stealing the models by using specially curated queries against the target model. This task is well accomplished by using part of the training data or a surrogate dataset to train a new model that mimics a target model in a white-box environment. In pragmatic situations, however, the target models are trained on private datasets that are inaccessible to the adversary. The data-free model extraction technique replaces this problem when it comes to using queries artificially curated by a generator similar to that used in Generative Adversarial Nets. We propose for the first time, to the best of our knowledge, an adversary black box attack extending to a regression problem for predicting bounding box coordinates in object detection. As part of our study, we found that defining a loss function and using a novel generator setup is one of the key aspects in extracting the target model. We find that the proposed model extraction method achieves significant results by using reasonable queries. The discovery of this object detection vulnerability will support future prospects for securing such models.
Authors: André Peter Kelm, Lucas Schmidt, Tim Rolff, Christian Wilms, Ehsan Yaghoubi, Simone Frintrop
In this work, we parallelize high-level features in deep networks to selectively skip or select class-specific features to reduce inference costs. This challenges most deep learning methods due to their limited ability to efficiently and effectively focus on selected class-specific features without retraining. We propose a serial-parallel hybrid architecture with serial generic low-level features and parallel high-level features. This accounts for the fact that many high-level features are class-specific rather than generic, and has connections to recent neuroscientific findings that observe spatially and contextually separated neural activations in the human brain. Our approach provides the unique functionality of cutouts: selecting parts of the network to focus on only relevant subsets of classes without requiring retraining. High performance is maintained, but the cost of inference can be significantly reduced. In some of our examples, up to $75\,\%$ of parameters are skipped and $35\,\%$ fewer GMACs (Giga multiply-accumulate) operations are used as the approach adapts to a change in task complexity. This is important for mobile, industrial, and robotic applications where reducing the number of parameters, the computational complexity, and thus the power consumption can be paramount. Another unique functionality is that it allows processing to be directly influenced by enhancing or inhibiting high-level class-specific features, similar to the mechanism of selective attention in the human brain. This can be relevant for cross-modal applications, the use of semantic prior knowledge, and/or context-aware processing.
Authors: Nina Weng, Siavash Bigdeli, Eike Petersen, Aasa Feragen
While many studies have assessed the fairness of AI algorithms in the medical field, the causes of differences in prediction performance are often unknown. This lack of knowledge about the causes of bias hampers the efficacy of bias mitigation, as evidenced by the fact that simple dataset balancing still often performs best in reducing performance gaps but is unable to resolve all performance differences. In this work, we investigate the causes of gender bias in machine learning-based chest X-ray diagnosis. In particular, we explore the hypothesis that breast tissue leads to underexposure of the lungs and causes lower model performance. Methodologically, we propose a new sampling method which addresses the highly skewed distribution of recordings per patient in two widely used public datasets, while at the same time reducing the impact of label errors. Our comprehensive analysis of gender differences across diseases, datasets, and gender representations in the training set shows that dataset imbalance is not the sole cause of performance differences. Moreover, relative group performance differs strongly between datasets, indicating important dataset-specific factors influencing male/female group performance. Finally, we investigate the effect of breast tissue more specifically, by cropping out the breasts from recordings, finding that this does not resolve the observed performance gaps. In conclusion, our results indicate that dataset-specific factors, not fundamental physiological differences, are the main drivers of male--female performance gaps in chest X-ray analyses on widely used NIH and CheXpert Dataset.
Authors: Fan Bai, Xiaohan Xing, Yutian Shen, Han Ma, Max Q.-H. Meng
Weakly supervised methods, such as class activation maps (CAM) based, have been applied to achieve bleeding segmentation with low annotation efforts in Wireless Capsule Endoscopy (WCE) images. However, the CAM labels tend to be extremely noisy, and there is an irreparable gap between CAM labels and ground truths for medical images. This paper proposes a new Discrepancy-basEd Active Learning (DEAL) approach to bridge the gap between CAMs and ground truths with a few annotations. Specifically, to liberate labor, we design a novel discrepancy decoder model and a CAMPUS (CAM, Pseudo-label and groUnd-truth Selection) criterion to replace the noisy CAMs with accurate model predictions and a few human labels. The discrepancy decoder model is trained with a unique scheme to generate standard, coarse and fine predictions. And the CAMPUS criterion is proposed to predict the gaps between CAMs and ground truths based on model divergence and CAM divergence. We evaluate our method on the WCE dataset and results show that our method outperforms the state-of-the-art active learning methods and reaches comparable performance to those trained with full annotated datasets with only 10% of the training data labeled.
Authors: Yidong Cai, Jie Liu, Jie Tang, Gangshan Wu
Object modeling has become a core part of recent tracking frameworks. Current popular tackers use Transformer attention to extract the template feature separately or interactively with the search region. However, separate template learning lacks communication between the template and search regions, which brings difficulty in extracting discriminative target-oriented features. On the other hand, interactive template learning produces hybrid template features, which may introduce potential distractors to the template via the cluttered search regions. To enjoy the merits of both methods, we propose a robust object modeling framework for visual tracking (ROMTrack), which simultaneously models the inherent template and the hybrid template features. As a result, harmful distractors can be suppressed by combining the inherent features of target objects with search regions' guidance. Target-related features can also be extracted using the hybrid template, thus resulting in a more robust object modeling framework. To further enhance robustness, we present novel variation tokens to depict the ever-changing appearance of target objects. Variation tokens are adaptable to object deformation and appearance variations, which can boost overall performance with negligible computation. Experiments show that our ROMTrack sets a new state-of-the-art on multiple benchmarks.
Authors: Nikhel Gupta, Zeeshan Hayder, Ray P. Norris, Minh Huynh, Lars Petersson, X. Rosalind Wang, Heinz Andernach, Bärbel S. Koribalski, Miranda Yew, Evan J. Crawford
The present work discusses the use of a weakly-supervised deep learning algorithm that reduces the cost of labelling pixel-level masks for complex radio galaxies with multiple components. The algorithm is trained on weak class-level labels of radio galaxies to get class activation maps (CAMs). The CAMs are further refined using an inter-pixel relations network (IRNet) to get instance segmentation masks over radio galaxies and the positions of their infrared hosts. We use data from the Australian Square Kilometre Array Pathfinder (ASKAP) telescope, specifically the Evolutionary Map of the Universe (EMU) Pilot Survey, which covered a sky area of 270 square degrees with an RMS sensitivity of 25-35 $\mu$Jy/beam. We demonstrate that weakly-supervised deep learning algorithms can achieve high accuracy in predicting pixel-level information, including masks for the extended radio emission encapsulating all galaxy components and the positions of the infrared host galaxies. We evaluate the performance of our method using mean Average Precision (mAP) across multiple classes at a standard intersection over union (IoU) threshold of 0.5. We show that the model achieves a mAP$_{50}$ of 67.5\% and 76.8\% for radio masks and infrared host positions, respectively. The network architecture can be found at the following link: https://github.com/Nikhel1/Gal-CAM
Authors: Changjian Chen, Yukai Guo, Fengyuan Tian, Shilong Liu, Weikai Yang, Zhaowei Wang, Jing Wu, Hang Su, Hanspeter Pfister, Shixia Liu
Existing model evaluation tools mainly focus on evaluating classification models, leaving a gap in evaluating more complex models, such as object detection. In this paper, we develop an open-source visual analysis tool, Uni-Evaluator, to support a unified model evaluation for classification, object detection, and instance segmentation in computer vision. The key idea behind our method is to formulate both discrete and continuous predictions in different tasks as unified probability distributions. Based on these distributions, we develop 1) a matrix-based visualization to provide an overview of model performance; 2) a table visualization to identify the problematic data subsets where the model performs poorly; 3) a grid visualization to display the samples of interest. These visualizations work together to facilitate the model evaluation from a global overview to individual samples. Two case studies demonstrate the effectiveness of Uni-Evaluator in evaluating model performance and making informed improvements.
Authors: Md. Simul Hasan Talukder, Ajay Kirshno Sarkar, Sharmin Akter, Md. Nuhi-Alamin
Diabetic Retinopathy (DR) is an ocular condition caused by a sustained high level of sugar in the blood, which causes the retinal capillaries to block and bleed, causing retinal tissue damage. It usually results in blindness. Early detection can help in lowering the risk of DR and its severity. The robust and accurate prediction and detection of diabetic retinopathy is a challenging task. This paper develops a machine learning model for detecting Diabetic Retinopathy that is entirely accurate. Pre-trained models such as ResNet50, InceptionV3, Xception, DenseNet121, VGG19, NASNetMobile, MobileNetV2, DensNet169, and DenseNet201 with pooling layer, dense layer, and appropriate dropout layer at the bottom of them were carried out in transfer learning (TL) approach. Data augmentation and regularization was performed to reduce overfitting. Transfer Learning model of DenseNet121, Average and weighted ensemble of DenseNet169 and DenseNet201 TL architectures contribute individually the highest accuracy of 100%, the highest precision, recall, F-1 score of 100%, 100%, and 100%, respectively.
Authors: Md. Simul Hasan Talukder, Mohammad Raziuddin Chowdhury, Md Sakib Ullah Sourav, Abdullah Al Rakin, Shabbir Ahmed Shuvo, Rejwan Bin Sulaiman, Musarrat Saberin Nipun, Muntarin Islam, Mst Rumpa Islam, Md Aminul Islam, Zubaer Haque
In certain Asian countries, Jute is one of the primary sources of income and Gross Domestic Product (GDP) for the agricultural sector. Like many other crops, Jute is prone to pest infestations, and its identification is typically made visually in countries like Bangladesh, India, Myanmar, and China. In addition, this method is time-consuming, challenging, and somewhat imprecise, which poses a substantial financial risk. To address this issue, the study proposes a high-performing and resilient transfer learning (TL) based JutePestDetect model to identify jute pests at the early stage. Firstly, we prepared jute pest dataset containing 17 classes and around 380 photos per pest class, which were evaluated after manual and automatic pre-processing and cleaning, such as background removal and resizing. Subsequently, five prominent pre-trained models -DenseNet201, InceptionV3, MobileNetV2, VGG19, and ResNet50 were selected from a previous study to design the JutePestDetect model. Each model was revised by replacing the classification layer with a global average pooling layer and incorporating a dropout layer for regularization. To evaluate the models performance, various metrics such as precision, recall, F1 score, ROC curve, and confusion matrix were employed. These analyses provided additional insights for determining the efficacy of the models. Among them, the customized regularized DenseNet201-based proposed JutePestDetect model outperformed the others, achieving an impressive accuracy of 99%. As a result, our proposed method and strategy offer an enhanced approach to pest identification in the case of Jute, which can significantly benefit farmers worldwide.
Authors: Miguel-Ángel Fernández-Torres
This PhD. Thesis concerns the study and development of hierarchical representations for spatio-temporal visual attention modeling and understanding in video sequences. More specifically, we propose two computational models for visual attention. First, we present a generative probabilistic model for context-aware visual attention modeling and understanding. Secondly, we develop a deep network architecture for visual attention modeling, which first estimates top-down spatio-temporal visual attention, and ultimately serves for modeling attention in the temporal domain.
Authors: Meng Wei, Charlie Budd, Luis C. Garcia-Peraza-Herrera, Reuben Dorent, Miaojing Shi, Tom Vercauteren
Surgical instrument segmentation is recognised as a key enabler to provide advanced surgical assistance and improve computer assisted interventions. In this work, we propose SegMatch, a semi supervised learning method to reduce the need for expensive annotation for laparoscopic and robotic surgical images. SegMatch builds on FixMatch, a widespread semi supervised classification pipeline combining consistency regularization and pseudo labelling, and adapts it for the purpose of segmentation. In our proposed SegMatch, the unlabelled images are weakly augmented and fed into the segmentation model to generate a pseudo-label to enforce the unsupervised loss against the output of the model for the adversarial augmented image on the pixels with a high confidence score. Our adaptation for segmentation tasks includes carefully considering the equivariance and invariance properties of the augmentation functions we rely on. To increase the relevance of our augmentations, we depart from using only handcrafted augmentations and introduce a trainable adversarial augmentation strategy. Our algorithm was evaluated on the MICCAI Instrument Segmentation Challenge datasets Robust-MIS 2019 and EndoVis 2017. Our results demonstrate that adding unlabelled data for training purposes allows us to surpass the performance of fully supervised approaches which are limited by the availability of training data in these challenges. SegMatch also outperforms a range of state-of-the-art semi-supervised learning semantic segmentation models in different labelled to unlabelled data ratios.
Authors: Faisal Hawlader, François Robinet, Raphaël Frank
Environmental perception is a key element of autonomous driving because the information received from the perception module influences core driving decisions. An outstanding challenge in real-time perception for autonomous driving lies in finding the best trade-off between detection quality and latency. Major constraints on both computation and power have to be taken into account for real-time perception in autonomous vehicles. Larger object detection models tend to produce the best results, but are also slower at runtime. Since the most accurate detectors cannot run in real-time locally, we investigate the possibility of offloading computation to edge and cloud platforms, which are less resource-constrained. We create a synthetic dataset to train object detection models and evaluate different offloading strategies. Using real hardware and network simulations, we compare different trade-offs between prediction quality and end-to-end delay. Since sending raw frames over the network implies additional transmission delays, we also explore the use of JPEG and H.265 compression at varying qualities and measure their impact on prediction metrics. We show that models with adequate compression can be run in real-time on the cloud while outperforming local detection performance.
Authors: Ali Jamali, Swalpa Kumar Roy, Danfeng Hong, Peter M Atkinson, Pedram Ghamisi
Convolutional Neural Networks (CNNs) are models that are utilized extensively for the hierarchical extraction of features. Vision transformers (ViTs), through the use of a self-attention mechanism, have recently achieved superior modeling of global contextual information compared to CNNs. However, to realize their image classification strength, ViTs require substantial training datasets. Where the available training data are limited, current advanced multi-layer perceptrons (MLPs) can provide viable alternatives to both deep CNNs and ViTs. In this paper, we developed the SGU-MLP, a learning algorithm that effectively uses both MLPs and spatial gating units (SGUs) for precise land use land cover (LULC) mapping. Results illustrated the superiority of the developed SGU-MLP classification algorithm over several CNN and CNN-ViT-based models, including HybridSN, ResNet, iFormer, EfficientFormer and CoAtNet. The proposed SGU-MLP algorithm was tested through three experiments in Houston, USA, Berlin, Germany and Augsburg, Germany. The SGU-MLP classification model was found to consistently outperform the benchmark CNN and CNN-ViT-based algorithms. For example, for the Houston experiment, SGU-MLP significantly outperformed HybridSN, CoAtNet, Efficientformer, iFormer and ResNet by approximately 15%, 19%, 20%, 21%, and 25%, respectively, in terms of average accuracy. The code will be made publicly available at https://github.com/aj1365/SGUMLP
Authors: Luv Verma, Varun Mohan
This study performs an ablation analysis of Vector Quantized Generative Adversarial Networks (VQGANs), concentrating on image-to-image synthesis utilizing a single NVIDIA A100 GPU. The current work explores the nuanced effects of varying critical parameters including the number of epochs, image count, and attributes of codebook vectors and latent dimensions, specifically within the constraint of limited resources. Notably, our focus is pinpointed on the vector quantization loss, keeping other hyperparameters and loss components (GAN loss) fixed. This was done to delve into a deeper understanding of the discrete latent space, and to explore how varying its size affects the reconstruction. Though, our results do not surpass the existing benchmarks, however, our findings shed significant light on VQGAN's behaviour for a smaller dataset, particularly concerning artifacts, codebook size optimization, and comparative analysis with Principal Component Analysis (PCA). The study also uncovers the promising direction by introducing 2D positional encodings, revealing a marked reduction in artifacts and insights into balancing clarity and overfitting.
Authors: Xumin Gao, Wenxin Xue, Callum Lennox, Mark Stevens, Junfeng Gao
Aphids are efficient vectors to transmit virus yellows in sugar beet fields. Timely monitoring and control of their populations are thus critical to prevent the large-scale outbreak of virus yellows. However, the manual counting of aphids, which is the most common practice, is labor-intensive and time-consuming. Additionally, two of the biggest challenges in aphid counting are that aphids are small objects and their density distributions are varied in different areas of the field. To address these challenges, we proposed a hybrid automatic aphid counting network architecture which integrates the detection network and the density map estimation network. When the distribution density of aphids is low, it utilizes an improved Yolov5 to count aphids. Conversely, when the distribution density of aphids is high, its witches to CSRNet to count aphids. To the best of our knowledge, this is the first framework integrating the detection network and the density map estimation network for counting tasks. Through comparison experiments of counting aphids, it verified that our proposed approach outperforms all other methods in counting aphids. It achieved the lowest MAE and RMSE values for both the standard and high-density aphid datasets: 2.93 and 4.01 (standard), and 34.19 and 38.66 (high-density), respectively. Moreover, the AP of the improved Yolov5 is 5% higher than that of the original Yolov5. Especially for extremely small aphids and densely distributed aphids, the detection performance of the improved Yolov5 is significantly better than the original Yolov5. This work provides an effective early warning for the virus yellows risk caused by aphids in sugar beet fields, offering protection for sugar beet growth and ensuring sugar beet yield. The datasets and project code are released at: https://github.com/JunfengGaolab/Counting-Aphids.
Authors: Soumyaroop Nandi, Prem Natarajan, Wael Abd-Almageed
The evaluation datasets and metrics for image manipulation detection and localization (IMDL) research have been standardized. But the training dataset for such a task is still nonstandard. Previous researchers have used unconventional and deviating datasets to train neural networks for detecting image forgeries and localizing pixel maps of manipulated regions. For a fair comparison, the training set, test set, and evaluation metrics should be persistent. Hence, comparing the existing methods may not seem fair as the results depend heavily on the training datasets as well as the model architecture. Moreover, none of the previous works release the synthetic training dataset used for the IMDL task. We propose a standardized benchmark training dataset for image splicing, copy-move forgery, removal forgery, and image enhancement forgery. Furthermore, we identify the problems with the existing IMDL datasets and propose the required modifications. We also train the state-of-the-art IMDL methods on our proposed TrainFors1 dataset for a fair evaluation and report the actual performance of these methods under similar conditions.
Authors: Xinyu Lyu, Jingwei Liu, Yuyu Guo, Lianli Gao
The task of dynamic scene graph generation (DynSGG) aims to generate scene graphs for given videos, which involves modeling the spatial-temporal information in the video. However, due to the long-tailed distribution of samples in the dataset, previous DynSGG models fail to predict the tail predicates. We argue that this phenomenon is due to previous methods that only pay attention to the local spatial-temporal information and neglect the consistency of multiple frames. To solve this problem, we propose a novel DynSGG model based on multi-task learning, DynSGG-MTL, which introduces the local interaction information and global human-action interaction information. The interaction between objects and frame features makes the model more fully understand the visual context of the single image. Long-temporal human actions supervise the model to generate multiple scene graphs that conform to the global constraints and avoid the model being unable to learn the tail predicates. Extensive experiments on Action Genome dataset demonstrate the efficacy of our proposed framework, which not only improves the dynamic scene graph generation but also alleviates the long-tail problem.
Authors: Lianli Gao, Xinyu Lyu, Yuyu Guo, Yuxuan Hu, Yuan-Fang Li, Lu Xu, Heng Tao Shen, Jingkuan Song
Scene graph generation aims to detect visual relationship triplets, (subject, predicate, object). Due to biases in data, current models tend to predict common predicates, e.g. "on" and "at", instead of informative ones, e.g. "standing on" and "looking at". This tendency results in the loss of precise information and overall performance. If a model only uses "stone on road" rather than "stone blocking road" to describe an image, it may be a grave misunderstanding. We argue that this phenomenon is caused by two imbalances: semantic space level imbalance and training sample level imbalance. For this problem, we propose DB-SGG, an effective framework based on debiasing but not the conventional distribution fitting. It integrates two components: Semantic Debiasing (SD) and Balanced Predicate Learning (BPL), for these imbalances. SD utilizes a confusion matrix and a bipartite graph to construct predicate relationships. BPL adopts a random undersampling strategy and an ambiguity removing strategy to focus on informative predicates. Benefiting from the model-agnostic process, our method can be easily applied to SGG models and outperforms Transformer by 136.3%, 119.5%, and 122.6% on mR@20 at three SGG sub-tasks on the SGG-VG dataset. Our method is further verified on another complex SGG dataset (SGG-GQA) and two downstream tasks (sentence-to-graph retrieval and image captioning).
Authors: Hongbo Kang, Yong Wang, Mengyuan Liu, Doudou Wu, Peng Liu, Wenming Yang
Reconstructing 3D poses from 2D poses lacking depth information is particularly challenging due to the complexity and diversity of human motion. The key is to effectively model the spatial constraints between joints to leverage their inherent dependencies. Thus, we propose a novel model, called Double-chain Graph Convolutional Transformer (DC-GCT), to constrain the pose through a double-chain design consisting of local-to-global and global-to-local chains to obtain a complex representation more suitable for the current human pose. Specifically, we combine the advantages of GCN and Transformer and design a Local Constraint Module (LCM) based on GCN and a Global Constraint Module (GCM) based on self-attention mechanism as well as a Feature Interaction Module (FIM). The proposed method fully captures the multi-level dependencies between human body joints to optimize the modeling capability of the model. Moreover, we propose a method to use temporal information into the single-frame model by guiding the video sequence embedding through the joint embedding of the target frame, with negligible increase in computational cost. Experimental results demonstrate that DC-GCT achieves state-of-the-art performance on two challenging datasets (Human3.6M and MPI-INF-3DHP). Notably, our model achieves state-of-the-art performance on all action categories in the Human3.6M dataset using detected 2D poses from CPN, and our code is available at: https://github.com/KHB1698/DC-GCT.
Authors: Yulin Yang, Patrick Geneva, Guoquan Huang
In this paper, we study state estimation of multi-visual-inertial systems (MVIS) and develop sensor fusion algorithms to optimally fuse an arbitrary number of asynchronous inertial measurement units (IMUs) or gyroscopes and global and(or) rolling shutter cameras. We are especially interested in the full calibration of the associated visual-inertial sensors, including the IMU or camera intrinsics and the IMU-IMU(or camera) spatiotemporal extrinsics as well as the image readout time of rolling-shutter cameras (if used). To this end, we develop a new analytic combined IMU integration with intrinsics-termed ACI3-to preintegrate IMU measurements, which is leveraged to fuse auxiliary IMUs and(or) gyroscopes alongside a base IMU. We model the multi-inertial measurements to include all the necessary inertial intrinsic and IMU-IMU spatiotemporal extrinsic parameters, while leveraging IMU-IMU rigid-body constraints to eliminate the necessity of auxiliary inertial poses and thus reducing computational complexity. By performing observability analysis of MVIS, we prove that the standard four unobservable directions remain - no matter how many inertial sensors are used, and also identify, for the first time, degenerate motions for IMU-IMU spatiotemporal extrinsics and auxiliary inertial intrinsics. In addition to the extensive simulations that validate our analysis and algorithms, we have built our own MVIS sensor rig and collected over 25 real-world datasets to experimentally verify the proposed calibration against the state-of-the-art calibration method such as Kalibr. We show that the proposed MVIS calibration is able to achieve competing accuracy with improved convergence and repeatability, which is open sourced to better benefit the community.
Authors: Wenjian Yao, Jiajun Bai, Wei Liao, Yuheng Chen, Mengjuan Liu, Yao Xie
Medical image segmentation is an important step in medical image analysis, especially as a crucial prerequisite for efficient disease diagnosis and treatment. The use of deep learning for image segmentation has become a prevalent trend. The widely adopted approach currently is U-Net and its variants. Additionally, with the remarkable success of pre-trained models in natural language processing tasks, transformer-based models like TransUNet have achieved desirable performance on multiple medical image segmentation datasets. In this paper, we conduct a survey of the most representative four medical image segmentation models in recent years. We theoretically analyze the characteristics of these models and quantitatively evaluate their performance on two benchmark datasets (i.e., Tuberculosis Chest X-rays and ovarian tumors). Finally, we discuss the main challenges and future trends in medical image segmentation. Our work can assist researchers in the related field to quickly establish medical segmentation models tailored to specific regions.
Authors: Huilin Zhu, Jingling Yuan, Xian Zhong, Zhengwei Yang, Zheng Wang, Shengfeng He
Domain adaptation is commonly employed in crowd counting to bridge the domain gaps between different datasets. However, existing domain adaptation methods tend to focus on inter-dataset differences while overlooking the intra-differences within the same dataset, leading to additional learning ambiguities. These domain-agnostic factors, e.g., density, surveillance perspective, and scale, can cause significant in-domain variations, and the misalignment of these factors across domains can lead to a drop in performance in cross-domain crowd counting. To address this issue, we propose a Domain-agnostically Aligned Optimal Transport (DAOT) strategy that aligns domain-agnostic factors between domains. The DAOT consists of three steps. First, individual-level differences in domain-agnostic factors are measured using structural similarity (SSIM). Second, the optimal transfer (OT) strategy is employed to smooth out these differences and find the optimal domain-to-domain misalignment, with outlier individuals removed via a virtual "dustbin" column. Third, knowledge is transferred based on the aligned domain-agnostic factors, and the model is retrained for domain adaptation to bridge the gap across domains. We conduct extensive experiments on five standard crowd-counting benchmarks and demonstrate that the proposed method has strong generalizability across diverse datasets. Our code will be available at: https://github.com/HopooLinZ/DAOT/.
Authors: Shaocong Liu, Tao Wang, Yan Zhang, Ruqin Zhou, Li Li, Chenguang Dai, Yongsheng Zhang, Hanyun Wang
The current point cloud registration methods are mainly based on geometric information and usually ignore the semantic information in the point clouds. In this paper, we treat the point cloud registration problem as semantic instance matching and registration task, and propose a deep semantic graph matching method for large-scale outdoor point cloud registration. Firstly, the semantic category labels of 3D point clouds are obtained by utilizing large-scale point cloud semantic segmentation network. The adjacent points with the same category labels are then clustered together by using Euclidean clustering algorithm to obtain the semantic instances. Secondly, the semantic adjacency graph is constructed based on the spatial adjacency relation of semantic instances. Three kinds of high-dimensional features including geometric shape features, semantic categorical features and spatial distribution features are learned through graph convolutional network, and enhanced based on attention mechanism. Thirdly, the semantic instance matching problem is modeled as an optimal transport problem, and solved through an optimal matching layer. Finally, according to the matched semantic instances, the geometric transformation matrix between two point clouds is first obtained by SVD algorithm and then refined by ICP algorithm. The experiments are cconducted on the KITTI Odometry dataset, and the average relative translation error and average relative rotation error of the proposed method are 6.6cm and 0.229{\deg} respectively.
Authors: Chang Nie, Guangming Wang, Zhe Liu, Luca Cavalli, Marc Pollefeys, Hesheng Wang
Robust estimation is a crucial and still challenging task, which involves estimating model parameters in noisy environments. Although conventional sampling consensus-based algorithms sample several times to achieve robustness, these algorithms cannot use data features and historical information effectively. In this paper, we propose RLSAC, a novel Reinforcement Learning enhanced SAmple Consensus framework for end-to-end robust estimation. RLSAC employs a graph neural network to utilize both data and memory features to guide exploring directions for sampling the next minimum set. The feedback of downstream tasks serves as the reward for unsupervised training. Therefore, RLSAC can avoid differentiating to learn the features and the feedback of downstream tasks for end-to-end robust estimation. In addition, RLSAC integrates a state transition module that encodes both data and memory features. Our experimental results demonstrate that RLSAC can learn from features to gradually explore a better hypothesis. Through analysis, it is apparent that RLSAC can be easily transferred to other sampling consensus-based robust estimation tasks. To the best of our knowledge, RLSAC is also the first method that uses reinforcement learning to sample consensus for end-to-end robust estimation. We release our codes at https://github.com/IRMVLab/RLSAC.
Authors: Yanjie Li, Mingxing Duan, Bin Xiao
The rudimentary adversarial attacks utilize additive noise to attack facial recognition (FR) models. However, because manipulating the total face is impractical in the physical setting, most real-world FR attacks are based on adversarial patches, which limit perturbations to a small area. Previous adversarial patch attacks often resulted in unnatural patterns and clear boundaries that were easily noticeable. In this paper, we argue that generating adversarial patches with plausible content can result in stronger transferability than using additive noise or directly sampling from the latent space. To generate natural-looking and highly transferable adversarial patches, we propose an innovative two-stage coarse-to-fine attack framework called Adv-Inpainting. In the first stage, we propose an attention-guided StyleGAN (Att-StyleGAN) that adaptively combines texture and identity features based on the attention map to generate high-transferable and natural adversarial patches. In the second stage, we design a refinement network with a new boundary variance loss to further improve the coherence between the patch and its surrounding area. Experiment results demonstrate that Adv-Inpainting is stealthy and can produce adversarial patches with stronger transferability and improved visual quality than previous adversarial patch attacks.
Authors: Alvaro Fernandez-Quilez, Tobias Nordström, Fredrik Jäderling, Svein Reidar Kjosavik, Martin Eklund
Background: Prostate cancer (PC) MRI-based risk calculators are commonly based on biological (e.g. PSA), MRI markers (e.g. volume), and patient age. Whilst patient age measures the amount of years an individual has existed, biological age (BA) might better reflect the physiology of an individual. However, surrogates from prostate MRI and linkage with clinically significant PC (csPC) remain to be explored. Purpose: To obtain and evaluate Prostate Age Gap (PAG) as an MRI marker tool for csPC risk. Study type: Retrospective. Population: A total of 7243 prostate MRI slices from 468 participants who had undergone prostate biopsies. A deep learning model was trained on 3223 MRI slices cropped around the gland from 81 low-grade PC (ncsPC, Gleason score <=6) and 131 negative cases and tested on the remaining 256 participants. Assessment: Chronological age was defined as the age of the participant at the time of the visit and used to train the deep learning model to predict the age of the patient. Following, we obtained PAG, defined as the model predicted age minus the patient's chronological age. Multivariate logistic regression models were used to estimate the association through odds ratio (OR) and predictive value of PAG and compared against PSA levels and PI-RADS>=3. Statistical tests: T-test, Mann-Whitney U test, Permutation test and ROC curve analysis. Results: The multivariate adjusted model showed a significant difference in the odds of clinically significant PC (csPC, Gleason score >=7) (OR =3.78, 95% confidence interval (CI):2.32-6.16, P <.001). PAG showed a better predictive ability when compared to PI-RADS>=3 and adjusted by other risk factors, including PSA levels: AUC =0.981 vs AUC =0.704, p<.001. Conclusion: PAG was significantly associated with the risk of clinically significant PC and outperformed other well-established PC risk factors.
Authors: Defang Cai, Pan Mu, Sixian Chan, Zhanpeng Shao, Cong Bai
As a common natural weather condition, rain can obscure video frames and thus affect the performance of the visual system, so video derain receives a lot of attention. In natural environments, rain has a wide variety of streak types, which increases the difficulty of the rain removal task. In this paper, we propose a Rain Review-based General video derain Network via knowledge distillation (named RRGNet) that handles different rain streak types with one pre-training weight. Specifically, we design a frame grouping-based encoder-decoder network that makes full use of the temporal information of the video. Further, we use the old task model to guide the current model in learning new rain streak types while avoiding forgetting. To consolidate the network's ability to derain, we design a rain review module to play back data from old tasks for the current model. The experimental results show that our developed general method achieves the best results in terms of running speed and derain effect.
Authors: Xinquan Yang, Jinheng Xie, Xuechen Li, Xuguang Li, Linlin Shen, Yongqiang Deng
In implant prosthesis treatment, the surgical guide of implant is used to ensure accurate implantation. However, such design heavily relies on the manual location of the implant position. When deep neural network has been proposed to assist the dentist in locating the implant position, most of them take a single slice as input, which do not fully explore 3D contextual information and ignoring the influence of implant slope. In this paper, we design a Text Guided 3D Context and Slope Aware Triple Network (TCSloT) which enables the perception of contextual information from multiple adjacent slices and awareness of variation of implant slopes. A Texture Variation Perception (TVP) module is correspondingly elaborated to process the multiple slices and capture the texture variation among slices and a Slope-Aware Loss (SAL) is proposed to dynamically assign varying weights for the regression head. Additionally, we design a conditional text guidance (CTG) module to integrate the text condition (i.e., left, middle and right) from the CLIP for assisting the implant position prediction. Extensive experiments on a dental implant dataset through five-fold cross-validation demonstrated that the proposed TCSloT achieves superior performance than existing methods.
Authors: Guozhang Liu, Baochai Peng, Ting Liu, Pan Zhang, Mengke Yuan, Chaoran Lu, Ningning Cao, Sen Zhang, Simin Huang, Tao Wang
The diversity of building architecture styles of global cities situated on various landforms, the degraded optical imagery affected by clouds and shadows, and the significant inter-class imbalance of roof types pose challenges for designing a robust and accurate building roof instance segmentor. To address these issues, we propose an effective framework to fulfill semantic interpretation of individual buildings with high-resolution optical satellite imagery. Specifically, the leveraged domain adapted pretraining strategy and composite dual-backbone greatly facilitates the discriminative feature learning. Moreover, new data augmentation pipeline, stochastic weight averaging (SWA) training and instance segmentation based model ensemble in testing are utilized to acquire additional performance boost. Experiment results show that our approach ranks in the first place of the 2023 IEEE GRSS Data Fusion Contest (DFC) Track 1 test phase ($mAP_{50}$:50.6\%). Note-worthily, we have also explored the potential of multimodal data fusion with both optical satellite imagery and SAR data.
Authors: Jie Hu, Chen Chen, Liujuan Cao, Shengchuan Zhang, Annan Shu, Guannan Jiang, Rongrong Ji
Pseudo-labeling is significant for semi-supervised instance segmentation, which generates instance masks and classes from unannotated images for subsequent training. However, in existing pipelines, pseudo-labels that contain valuable information may be directly filtered out due to mismatches in class and mask quality. To address this issue, we propose a novel framework, called pseudo-label aligning instance segmentation (PAIS), in this paper. In PAIS, we devise a dynamic aligning loss (DALoss) that adjusts the weights of semi-supervised loss terms with varying class and mask score pairs. Through extensive experiments conducted on the COCO and Cityscapes datasets, we demonstrate that PAIS is a promising framework for semi-supervised instance segmentation, particularly in cases where labeled data is severely limited. Notably, with just 1\% labeled data, PAIS achieves 21.2 mAP (based on Mask-RCNN) and 19.9 mAP (based on K-Net) on the COCO dataset, outperforming the current state-of-the-art model, \ie, NoisyBoundary with 7.7 mAP, by a margin of over 12 points. Code is available at: \url{https://github.com/hujiecpp/PAIS}.
Authors: Jiaqi Cui, Pinxian Zeng, Xinyi Zeng, Peng Wang, Xi Wu, Jiliu Zhou, Yan Wang, Dinggang Shen
To obtain high-quality positron emission tomography (PET) images while minimizing radiation exposure, various methods have been proposed for reconstructing standard-dose PET (SPET) images from low-dose PET (LPET) sinograms directly. However, current methods often neglect boundaries during sinogram-to-image reconstruction, resulting in high-frequency distortion in the frequency domain and diminished or fuzzy edges in the reconstructed images. Furthermore, the convolutional architectures, which are commonly used, lack the ability to model long-range non-local interactions, potentially leading to inaccurate representations of global structures. To alleviate these problems, we propose a transformer-based model that unites triple domains of sinogram, image, and frequency for direct PET reconstruction, namely TriDo-Former. Specifically, the TriDo-Former consists of two cascaded networks, i.e., a sinogram enhancement transformer (SE-Former) for denoising the input LPET sinograms and a spatial-spectral reconstruction transformer (SSR-Former) for reconstructing SPET images from the denoised sinograms. Different from the vanilla transformer that splits an image into 2D patches, based specifically on the PET imaging mechanism, our SE-Former divides the sinogram into 1D projection view angles to maintain its inner-structure while denoising, preventing the noise in the sinogram from prorogating into the image domain. Moreover, to mitigate high-frequency distortion and improve reconstruction details, we integrate global frequency parsers (GFPs) into SSR-Former. The GFP serves as a learnable frequency filter that globally adjusts the frequency components in the frequency domain, enforcing the network to restore high-frequency details resembling real SPET images. Validations on a clinical dataset demonstrate that our TriDo-Former outperforms the state-of-the-art methods qualitatively and quantitatively.
Authors: Tianchang Shen, Jacob Munkberg, Jon Hasselgren, Kangxue Yin, Zian Wang, Wenzheng Chen, Zan Gojcic, Sanja Fidler, Nicholas Sharp, Jun Gao
This work considers gradient-based mesh optimization, where we iteratively optimize for a 3D surface mesh by representing it as the isosurface of a scalar field, an increasingly common paradigm in applications including photogrammetry, generative modeling, and inverse physics. Existing implementations adapt classic isosurface extraction algorithms like Marching Cubes or Dual Contouring; these techniques were designed to extract meshes from fixed, known fields, and in the optimization setting they lack the degrees of freedom to represent high-quality feature-preserving meshes, or suffer from numerical instabilities. We introduce FlexiCubes, an isosurface representation specifically designed for optimizing an unknown mesh with respect to geometric, visual, or even physical objectives. Our main insight is to introduce additional carefully-chosen parameters into the representation, which allow local flexible adjustments to the extracted mesh geometry and connectivity. These parameters are updated along with the underlying scalar field via automatic differentiation when optimizing for a downstream task. We base our extraction scheme on Dual Marching Cubes for improved topological properties, and present extensions to optionally generate tetrahedral and hierarchically-adaptive meshes. Extensive experiments validate FlexiCubes on both synthetic benchmarks and real-world applications, showing that it offers significant improvements in mesh quality and geometric fidelity.
Authors: Chihiro Nakatani, Hiroaki Kawashima, Norimichi Ukita
This paper proposes joint attention estimation in a single image. Different from related work in which only the gaze-related attributes of people are independently employed, (I) their locations and actions are also employed as contextual cues for weighting their attributes, and (ii) interactions among all of these attributes are explicitly modeled in our method. For the interaction modeling, we propose a novel Transformer-based attention network to encode joint attention as low-dimensional features. We introduce a specialized MLP head with positional embedding to the Transformer so that it predicts pixelwise confidence of joint attention for generating the confidence heatmap. This pixelwise prediction improves the heatmap accuracy by avoiding the ill-posed problem in which the high-dimensional heatmap is predicted from the low-dimensional features. The estimated joint attention is further improved by being integrated with general image-based attention estimation. Our method outperforms SOTA methods quantitatively in comparative experiments. Code: https://anonymous.4open.science/r/anonymized_codes-ECA4.
Authors: Chaoran Lu, Ningning Cao, Pan Zhang, Ting Liu, Baochai Peng, Guozhang Liu, Mengke Yuan, Sen Zhang, Simin Huang, Tao Wang
Unifying the correlative single-view satellite image building extraction and height estimation tasks indicates a promising way to share representations and acquire generalist model for large-scale urban 3D reconstruction. However, the common spatial misalignment between building footprints and stereo-reconstructed nDSM height labels incurs degraded performance on both tasks. To address this issue, we propose a Height-hierarchy Guided Dual-decoder Network (HGDNet) to estimate building height. Under the guidance of synthesized discrete height-hierarchy nDSM, auxiliary height-hierarchical building extraction branch enhance the height estimation branch with implicit constraints, yielding an accuracy improvement of more than 6% on the DFC 2023 track2 dataset. Additional two-stage cascade architecture is adopted to achieve more accurate building extraction. Experiments on the DFC 2023 Track 2 dataset shows the superiority of the proposed method in building height estimation ({\delta}1:0.8012), instance extraction (AP50:0.7730), and the final average score 0.7871 ranks in the first place in test phase.
Authors: Sangeet Jaiswal, Dhruv Patel, Sreekanth Vempati, Konduru Saiswaroop
In a fashion e-commerce platform where customers can't physically examine the products on their own, being able to see other customers' text and image reviews of the product is critical while making purchase decisions. Given the high reliance on these reviews, over the years we have observed customers proactively sharing their reviews. With an increase in the coverage of User Generated Content (UGC), there has been a corresponding increase in the number of customer images. It is thus imperative to display the most relevant images on top as it may influence users' online shopping choices and behavior. In this paper, we propose a simple yet effective training procedure for ranking customer images. We created a dataset consisting of Myntra (A Major Indian Fashion e-commerce company) studio posts and highly engaged (upvotes/downvotes) UGC images as our starting point and used selected distortion techniques on the images of the above dataset to bring their quality at par with those of bad UGC images. We train our network to rank bad-quality images lower than high-quality ones. Our proposed method outperforms the baseline models on two metrics, namely correlation coefficient, and accuracy, by substantial margins.
Authors: Changkun Liu, Yukun Zhao, Tristan Braud
Visual Inertial Odometry (VIO) is an essential component of modern Augmented Reality (AR) applications. However, VIO only tracks the relative pose of the device, leading to drift over time. Absolute pose estimation methods infer the device's absolute pose, but their accuracy depends on the input quality. This paper introduces VIO-APR, a new framework for markerless mobile AR that combines an absolute pose regressor (APR) with a local VIO tracking system. VIO-APR uses VIO to assess the reliability of the APR and the APR to identify and compensate for VIO drift. This feedback loop results in more accurate positioning and more stable AR experiences. To evaluate VIO-APR, we created a dataset that combines camera images with ARKit's VIO system output for six indoor and outdoor scenes of various scales. Over this dataset, VIO-APR improves the median accuracy of popular APR by up to 36\% in position and 29\% in orientation, increases the percentage of frames in the high ($0.25 m, 2^{\circ}$) accuracy level by up to 112\% and reduces the percentage of frames predicted below the low ($5 m, 10^\circ$) accuracy greatly. We implement VIO-APR into a mobile AR application using Unity to demonstrate its capabilities. VIO-APR results in noticeably more accurate localization and a more stable overall experience.
Authors: Lanyun Zhu, Tianrun Chen, Jianxiong Yin, Simon See, Jun Liu
Extracting and using class-discriminative features is critical for fine-grained recognition. Existing works have demonstrated the possibility of applying deep CNNs to exploit features that distinguish similar classes. However, CNNs suffer from problems including frequency bias and loss of detailed local information, which restricts the performance of recognizing fine-grained categories. To address the challenge, we propose a novel texture branch as complimentary to the CNN branch for feature extraction. We innovatively utilize Gabor filters as a powerful extractor to exploit texture features, motivated by the capability of Gabor filters in effectively capturing multi-frequency features and detailed local information. We implement several designs to enhance the effectiveness of Gabor filters, including imposing constraints on parameter values and developing a learning method to determine the optimal parameters. Moreover, we introduce a statistical feature extractor to utilize informative statistical information from the signals captured by Gabor filters, and a gate selection mechanism to enable efficient computation by only considering qualified regions as input for texture extraction. Through the integration of features from the Gabor-filter-based texture branch and CNN-based semantic branch, we achieve comprehensive information extraction. We demonstrate the efficacy of our method on multiple datasets, including CUB-200-2011, NA-bird, Stanford Dogs, and GTOS-mobile. State-of-the-art performance is achieved using our approach.
Authors: Xianqiang Lyu, Junhui Hou
This paper presents a novel and interpretable end-to-end learning framework, called the deep compensation unfolding network (DCUNet), for restoring light field (LF) images captured under low-light conditions. DCUNet is designed with a multi-stage architecture that mimics the optimization process of solving an inverse imaging problem in a data-driven fashion. The framework uses the intermediate enhanced result to estimate the illumination map, which is then employed in the unfolding process to produce a new enhanced result. Additionally, DCUNet includes a content-associated deep compensation module at each optimization stage to suppress noise and illumination map estimation errors. To properly mine and leverage the unique characteristics of LF images, this paper proposes a pseudo-explicit feature interaction module that comprehensively exploits redundant information in LF images. The experimental results on both simulated and real datasets demonstrate the superiority of our DCUNet over state-of-the-art methods, both qualitatively and quantitatively. Moreover, DCUNet preserves the essential geometric structure of enhanced LF images much better. The code will be publicly available at https://github.com/lyuxianqiang/LFLL-DCU.
Authors: Francisco Mena, Diego Arenas, Marlon Nuske, Andreas Dengel
With a rapidly increasing amount and diversity of remote sensing (RS) data sources, there is a strong need for multi-view learning modeling. This is a complex task when considering the differences in resolution, magnitude, and noise of RS data. The typical approach for merging multiple RS sources has been input-level fusion, but other - more advanced - fusion strategies may outperform this traditional approach. This work assesses different fusion strategies for crop classification in the CropHarvest dataset. The fusion methods proposed in this work outperform models based on individual views and previous fusion methods. We do not find one single fusion method that consistently outperforms all other approaches. Instead, we present a comparison of multi-view fusion methods for three different datasets and show that, depending on the test region, different methods obtain the best performance. Despite this, we suggest a preliminary criterion for the selection of fusion methods.
Authors: Mohammad Zohaib, Alessio Del Bue
This paper proposes a new method to infer keypoints from arbitrary object categories in practical scenarios where point cloud data (PCD) are noisy, down-sampled and arbitrarily rotated. Our proposed model adheres to the following principles: i) keypoints inference is fully unsupervised (no annotation given), ii) keypoints position error should be low and resilient to PCD perturbations (robustness), iii) keypoints should not change their indexes for the intra-class objects (semantic coherence), iv) keypoints should be close to or proximal to PCD surface (compactness). We achieve these desiderata by proposing a new self-supervised training strategy for keypoints estimation that does not assume any a priori knowledge of the object class, and a model architecture with coupled auxiliary losses that promotes the desired keypoints properties. We compare the keypoints estimated by the proposed approach with those of the state-of-the-art unsupervised approaches. The experiments show that our approach outperforms by estimating keypoints with improved coverage (+9.41%) while being semantically consistent (+4.66%) that best characterizes the object's 3D shape for downstream tasks. Code and data are available at: https://github.com/IITPAVIS/SC3K
Authors: Guangyao Li, Wenxuan Hou, Di Hu
Audio-Visual Question Answering (AVQA) task aims to answer questions about different visual objects, sounds, and their associations in videos. Such naturally multi-modal videos are composed of rich and complex dynamic audio-visual components, where most of which could be unrelated to the given questions, or even play as interference in answering the content of interest. Oppositely, only focusing on the question-aware audio-visual content could get rid of influence, meanwhile enabling the model to answer more efficiently. In this paper, we propose a Progressive Spatio-Temporal Perception Network (PSTP-Net), which contains three modules that progressively identify key spatio-temporal regions w.r.t. questions. Specifically, a temporal segment selection module is first introduced to select the most relevant audio-visual segments related to the given question. Then, a spatial region selection module is utilized to choose the most relevant regions associated with the question from the selected temporal segments. To further refine the selection of features, an audio-guided visual attention module is employed to perceive the association between auido and selected spatial regions. Finally, the spatio-temporal features from these modules are integrated for answering the question. Extensive experimental results on the public MUSIC-AVQA and AVQA datasets provide compelling evidence of the effectiveness and efficiency of PSTP-Net. Code is available at: \href{https://github.com/GeWu-Lab/PSTP-Net}{https://github.com/GeWu-Lab/PSTP-Net}
Authors: Ruikai Cui, Siyuan He, Shi Qiu
Foundation models, such as OpenAI's GPT-3 and GPT-4, Meta's LLaMA, and Google's PaLM2, have revolutionized the field of artificial intelligence. A notable paradigm shift has been the advent of the Segment Anything Model (SAM), which has exhibited a remarkable capability to segment real-world objects, trained on 1 billion masks and 11 million images. Although SAM excels in general object segmentation, it lacks the intrinsic ability to detect salient objects, resulting in suboptimal performance in this domain. To address this challenge, we present the Segment Salient Object Model (SSOM), an innovative approach that adaptively fine-tunes SAM for salient object detection by harnessing the low-rank structure inherent in deep learning. Comprehensive qualitative and quantitative evaluations across five challenging RGB benchmark datasets demonstrate the superior performance of our approach, surpassing state-of-the-art methods.
Authors: Haozhe Wu, Songtao Zhou, Jia Jia, Junliang Xing, Qi Wen, Xiang Wen
Speech-driven 3D face animation poses significant challenges due to the intricacy and variability inherent in human facial movements. This paper emphasizes the importance of considering both the composite and regional natures of facial movements in speech-driven 3D face animation. The composite nature pertains to how speech-independent factors globally modulate speech-driven facial movements along the temporal dimension. Meanwhile, the regional nature alludes to the notion that facial movements are not globally correlated but are actuated by local musculature along the spatial dimension. It is thus indispensable to incorporate both natures for engendering vivid animation. To address the composite nature, we introduce an adaptive modulation module that employs arbitrary facial movements to dynamically adjust speech-driven facial movements across frames on a global scale. To accommodate the regional nature, our approach ensures that each constituent of the facial features for every frame focuses on the local spatial movements of 3D faces. Moreover, we present a non-autoregressive backbone for translating audio to 3D facial movements, which maintains high-frequency nuances of facial movements and facilitates efficient inference. Comprehensive experiments and user studies demonstrate that our method surpasses contemporary state-of-the-art approaches both qualitatively and quantitatively.
Authors: Jyoti Kini, Sarah Fleischer, Ishan Dave, Mubarak Shah
In this work, we propose an ensemble modeling approach for multimodal action recognition. We independently train individual modality models using a variant of focal loss tailored to handle the long-tailed distribution of the MECCANO [21] dataset. Based on the underlying principle of focal loss, which captures the relationship between tail (scarce) classes and their prediction difficulties, we propose an exponentially decaying variant of focal loss for our current task. It initially emphasizes learning from the hard misclassified examples and gradually adapts to the entire range of examples in the dataset. This annealing process encourages the model to strike a balance between focusing on the sparse set of hard samples, while still leveraging the information provided by the easier ones. Additionally, we opt for the late fusion strategy to combine the resultant probability distributions from RGB and Depth modalities for final action prediction. Experimental evaluations on the MECCANO dataset demonstrate the effectiveness of our approach.
Authors: Jun Zhou, Kai Chen, Linlin Xu, Qi Dou, Jing Qin
One critical challenge in 6D object pose estimation from a single RGBD image is efficient integration of two different modalities, i.e., color and depth. In this work, we tackle this problem by a novel Deep Fusion Transformer~(DFTr) block that can aggregate cross-modality features for improving pose estimation. Unlike existing fusion methods, the proposed DFTr can better model cross-modality semantic correlation by leveraging their semantic similarity, such that globally enhanced features from different modalities can be better integrated for improved information extraction. Moreover, to further improve robustness and efficiency, we introduce a novel weighted vector-wise voting algorithm that employs a non-iterative global optimization strategy for precise 3D keypoint localization while achieving near real-time inference. Extensive experiments show the effectiveness and strong generalization capability of our proposed 3D keypoint voting algorithm. Results on four widely used benchmarks also demonstrate that our method outperforms the state-of-the-art methods by large margins.
Authors: Hao Liang, Pietro Perona, Guha Balakrishnan
We propose an experimental method for measuring bias in face recognition systems. Existing methods to measure bias depend on benchmark datasets that are collected in the wild and annotated for protected (e.g., race, gender) and non-protected (e.g., pose, lighting) attributes. Such observational datasets only permit correlational conclusions, e.g., "Algorithm A's accuracy is different on female and male faces in dataset X.". By contrast, experimental methods manipulate attributes individually and thus permit causal conclusions, e.g., "Algorithm A's accuracy is affected by gender and skin color."
Our method is based on generating synthetic faces using a neural face generator, where each attribute of interest is modified independently while leaving all other attributes constant. Human observers crucially provide the ground truth on perceptual identity similarity between synthetic image pairs. We validate our method quantitatively by evaluating race and gender biases of three research-grade face recognition models. Our synthetic pipeline reveals that for these algorithms, accuracy is lower for Black and East Asian population subgroups. Our method can also quantify how perceptual changes in attributes affect face identity distances reported by these models. Our large synthetic dataset, consisting of 48,000 synthetic face image pairs (10,200 unique synthetic faces) and 555,000 human annotations (individual attributes and pairwise identity comparisons) is available to researchers in this important area.
Authors: Pan Mu, Hanning Xu, Zheyuan Liu, Zheng Wang, Sixian Chan, Cong Bai
Underwater images often suffer from color distortion and low contrast resulting in various image types, due to the scattering and absorption of light by water. While it is difficult to obtain high-quality paired training samples with a generalized model. To tackle these challenges, we design a Generalized Underwater image enhancement method via a Physical-knowledge-guided Dynamic Model (short for GUPDM), consisting of three parts: Atmosphere-based Dynamic Structure (ADS), Transmission-guided Dynamic Structure (TDS), and Prior-based Multi-scale Structure (PMS). In particular, to cover complex underwater scenes, this study changes the global atmosphere light and the transmission to simulate various underwater image types (e.g., the underwater image color ranging from yellow to blue) through the formation model. We then design ADS and TDS that use dynamic convolutions to adaptively extract prior information from underwater images and generate parameters for PMS. These two modules enable the network to select appropriate parameters for various water types adaptively. Besides, the multi-scale feature extraction module in PMS uses convolution blocks with different kernel sizes and obtains weights for each feature map via channel attention block and fuses them to boost the receptive field of the network. The source code will be available at \href{https://github.com/shiningZZ/GUPDM}{https://github.com/shiningZZ/GUPDM}.
Authors: Sahar Almahfouz Nasser, Ashutosh Sharma, Anmol Saraf, Amruta Mahendra Parulekar, Purvi Haria, Amit Sethi
Ultrasound (US) imaging is better suited for intraoperative settings because it is real-time and more portable than other imaging techniques, such as mammography. However, US images are characterized by lower spatial resolution noise-like artifacts. This research aims to address these limitations by providing surgeons with mammogram-like image quality in real-time from noisy US images. Unlike previous approaches for improving US image quality that aim to reduce artifacts by treating them as (speckle noise), we recognize their value as informative wave interference pattern (WIP). To achieve this, we utilize the Stride software to numerically solve the forward model, generating ultrasound images from mammograms images by solving wave-equations. Additionally, we leverage the power of domain adaptation to enhance the realism of the simulated ultrasound images. Then, we utilize generative adversarial networks (GANs) to tackle the inverse problem of generating mammogram-quality images from ultrasound images. The resultant images have considerably more discernible details than the original US images.
Authors: Changkun Liu, Yukun Zhao, Tristan Braud
Markerless Mobile Augmented Reality (AR) aims to anchor digital content in the physical world without using specific 2D or 3D objects. Absolute Pose Regressors (APR) are end-to-end machine learning solutions that infer the device's pose from a single monocular image. Thanks to their low computation cost, they can be directly executed on the constrained hardware of mobile AR devices. However, APR methods tend to yield significant inaccuracies for input images that are too distant from the training set. This paper introduces KS-APR, a pipeline that assesses the reliability of an estimated pose with minimal overhead by combining the inference results of the APR and the prior images in the training set. Mobile AR systems tend to rely upon visual-inertial odometry to track the relative pose of the device during the experience. As such, KS-APR favours reliability over frequency, discarding unreliable poses. This pipeline can integrate most existing APR methods to improve accuracy by filtering unreliable images with their pose estimates. We implement the pipeline on three types of APR models on indoor and outdoor datasets. The median error on position and orientation is reduced for all models, and the proportion of large errors is minimized across datasets. Our method enables state-of-the-art APRs such as DFNetdm to outperform single-image and sequential APR methods. These results demonstrate the scalability and effectiveness of KS-APR for visual localization tasks that do not require one-shot decisions.
Authors: Simon Dahan, Mariana da Silva, Daniel Rueckert, Emma C Robinson
Self-supervision has been widely explored as a means of addressing the lack of inductive biases in vision transformer architectures, which limits generalisation when networks are trained on small datasets. This is crucial in the context of cortical imaging, where phenotypes are complex and heterogeneous, but the available datasets are limited in size. This paper builds upon recent advancements in translating vision transformers to surface meshes and investigates the potential of Masked AutoEncoder (MAE) self-supervision for cortical surface learning. By reconstructing surface data from a masked version of the input, the proposed method effectively models cortical structure to learn strong representations that translate to improved performance in downstream tasks. We evaluate our approach on cortical phenotype regression using the developing Human Connectome Project (dHCP) and demonstrate that pre-training leads to a 26\% improvement in performance, with an 80\% faster convergence, compared to models trained from scratch. Furthermore, we establish that pre-training vision transformer models on large datasets, such as the UK Biobank (UKB), enables the acquisition of robust representations for finetuning in low-data scenarios. Our code and pre-trained models are publicly available at \url{https://github.com/metrics-lab/surface-vision-transformers}.
Authors: Patrick Palmer, Martin Krueger, Richard Altendorfer, Ganesh Adam, Torsten Bertram
Recent developments and the beginning market introduction of high-resolution imaging 4D (3+1D) radar sensors have initialized deep learning-based radar perception research. We investigate deep learning-based models operating on radar point clouds for 3D object detection. 3D object detection on lidar point cloud data is a mature area of 3D vision. Many different architectures have been proposed, each with strengths and weaknesses. Due to similarities between 3D lidar point clouds and 3+1D radar point clouds, those existing 3D object detectors are a natural basis to start deep learning-based 3D object detection on radar data. Thus, the first step is to analyze the detection performance of the existing models on the new data modality and evaluate them in depth. In order to apply existing 3D point cloud object detectors developed for lidar point clouds to the radar domain, they need to be adapted first. While some detectors, such as PointPillars, have already been adapted to be applicable to radar data, we have adapted others, e.g., Voxel R-CNN, SECOND, PointRCNN, and PV-RCNN. To this end, we conduct a cross-model validation (evaluating a set of models on one particular data set) as well as a cross-data set validation (evaluating all models in the model set on several data sets). The high-resolution radar data used are the View-of-Delft and Astyx data sets. Finally, we evaluate several adaptations of the models and their training procedures. We also discuss major factors influencing the detection performance on radar data and propose possible solutions indicating potential future research avenues.
Authors: Yuming Chen, Xinbin Yuan, Ruiqi Wu, Jiabao Wang, Qibin Hou, Ming-Ming Cheng
We aim at providing the object detection community with an efficient and performant object detector, termed YOLO-MS. The core design is based on a series of investigations on how convolutions with different kernel sizes affect the detection performance of objects at different scales. The outcome is a new strategy that can strongly enhance multi-scale feature representations of real-time object detectors. To verify the effectiveness of our strategy, we build a network architecture, termed YOLO-MS. We train our YOLO-MS on the MS COCO dataset from scratch without relying on any other large-scale datasets, like ImageNet, or pre-trained weights. Without bells and whistles, our YOLO-MS outperforms the recent state-of-the-art real-time object detectors, including YOLO-v7 and RTMDet, when using a comparable number of parameters and FLOPs. Taking the XS version of YOLO-MS as an example, with only 4.5M learnable parameters and 8.7G FLOPs, it can achieve an AP score of 43%+ on MS COCO, which is about 2%+ higher than RTMDet with the same model size. Moreover, our work can also be used as a plug-and-play module for other YOLO models. Typically, our method significantly improves the AP of YOLOv8 from 37%+ to 40%+ with even fewer parameters and FLOPs. Code is available at https://github.com/FishAndWasabi/YOLO-MS.
Authors: Xu Zheng, Tianbo Pan, Yunhao Luo, Lin Wang
Endeavors have been recently made to transfer knowledge from the labeled pinhole image domain to the unlabeled panoramic image domain via Unsupervised Domain Adaptation (UDA). The aim is to tackle the domain gaps caused by the style disparities and distortion problem from the non-uniformly distributed pixels of equirectangular projection (ERP). Previous works typically focus on transferring knowledge based on geometric priors with specially designed multi-branch network architectures. As a result, considerable computational costs are induced, and meanwhile, their generalization abilities are profoundly hindered by the variation of distortion among pixels. In this paper, we find that the pixels' neighborhood regions of the ERP indeed introduce less distortion. Intuitively, we propose a novel UDA framework that can effectively address the distortion problems for panoramic semantic segmentation. In comparison, our method is simpler, easier to implement, and more computationally efficient. Specifically, we propose distortion-aware attention (DA) capturing the neighboring pixel distribution without using any geometric constraints. Moreover, we propose a class-wise feature aggregation (CFA) module to iteratively update the feature representations with a memory bank. As such, the feature similarity between two domains can be consistently optimized. Extensive experiments show that our method achieves new state-of-the-art performance while remarkably reducing 80% parameters.
Authors: Meir Yossef Levi, Guy Gilboa
The ability to cope accurately and fast with Out-Of-Distribution (OOD) samples is crucial in real-world safety demanding applications. In this work we first study the interplay between critical points of 3D point clouds and OOD samples. Our findings are that common corruptions and outliers are often interpreted as critical points. We generalize the notion of critical points into importance measures. We show that training a classification network based only on less important points dramatically improves robustness, at a cost of minor performance loss on the clean set. We observe that normalized entropy is highly informative for corruption analysis. An adaptive threshold based on normalized entropy is suggested for selecting the set of uncritical points. Our proposed importance measure is extremely fast to compute. We show it can be used for a variety of applications, such as Explainable AI (XAI), Outlier Removal, Uncertainty Estimation, Robust Classification and Adversarial Defense. We reach SOTA results on the two latter tasks.
Authors: Frans de Boer, Jan C. van Gemert, Jouke Dijkstra, Silvia L. Pintea
Activity progress prediction aims to estimate what percentage of an activity has been completed. Currently this is done with machine learning approaches, trained and evaluated on complicated and realistic video datasets. The videos in these datasets vary drastically in length and appearance. And some of the activities have unanticipated developments, making activity progression difficult to estimate. In this work, we examine the results obtained by existing progress prediction methods on these datasets. We find that current progress prediction methods seem not to extract useful visual information for the progress prediction task. Therefore, these methods fail to exceed simple frame-counting baselines. We design a precisely controlled dataset for activity progress prediction and on this synthetic dataset we show that the considered methods can make use of the visual information, when this directly relates to the progress prediction. We conclude that the progress prediction task is ill-posed on the currently used real-world datasets. Moreover, to fairly measure activity progression we advise to consider a, simple but effective, frame-counting baseline.
Authors: Wongi Park, Inhyuk Park, Sungeun Kim, Jongbin Ryu
In real medical data, training samples typically show long-tailed distributions with multiple labels. Class distribution of the medical data has a long-tailed shape, in which the incidence of different diseases is quite varied, and at the same time, it is not unusual for images taken from symptomatic patients to be multi-label diseases. Therefore, in this paper, we concurrently address these two issues by putting forth a robust asymmetric loss on the polynomial function. Since our loss tackles both long-tailed and multi-label classification problems simultaneously, it leads to a complex design of the loss function with a large number of hyper-parameters. Although a model can be highly fine-tuned due to a large number of hyper-parameters, it is difficult to optimize all hyper-parameters at the same time, and there might be a risk of overfitting a model. Therefore, we regularize the loss function using the Hill loss approach, which is beneficial to be less sensitive against the numerous hyper-parameters so that it reduces the risk of overfitting the model. For this reason, the proposed loss is a generic method that can be applied to most medical image classification tasks and does not make the training process more time-consuming. We demonstrate that the proposed robust asymmetric loss performs favorably against the long-tailed with multi-label medical image classification in addition to the various long-tailed single-label datasets. Notably, our method achieves Top-5 results on the CXR-LT dataset of the ICCV CVAMD 2023 competition. We opensource our implementation of the robust asymmetric loss in the public repository: https://github.com/kalelpark/RAL.
Authors: Liang Chen, Jiawei Zhang, Zhenhua Li, Yunxuan Wei, Faming Fang, Jimmy Ren, Jinshan Pan
Images taken under the low-light condition often contain blur and saturated pixels at the same time. Deblurring images with saturated pixels is quite challenging. Because of the limited dynamic range, the saturated pixels are usually clipped in the imaging process and thus cannot be modeled by the linear blur model. Previous methods use manually designed smooth functions to approximate the clipping procedure. Their deblurring processes often require empirically defined parameters, which may not be the optimal choices for different images. In this paper, we develop a data-driven approach to model the saturated pixels by a learned latent map. Based on the new model, the non-blind deblurring task can be formulated into a maximum a posterior (MAP) problem, which can be effectively solved by iteratively computing the latent map and the latent image. Specifically, the latent map is computed by learning from a map estimation network (MEN), and the latent image estimation process is implemented by a Richardson-Lucy (RL)-based updating scheme. To estimate high-quality deblurred images without amplified artifacts, we develop a prior estimation network (PEN) to obtain prior information, which is further integrated into the RL scheme. Experimental results demonstrate that the proposed method performs favorably against state-of-the-art algorithms both quantitatively and qualitatively on synthetic and real-world images.
Authors: Xuehan Bai, Yan Li, Yanhua Cheng, Wenjie Yang, Quan Chen, Han Li
The proliferation of short video and live-streaming platforms has revolutionized how consumers engage in online shopping. Instead of browsing product pages, consumers are now turning to rich-content e-commerce, where they can purchase products through dynamic and interactive media like short videos and live streams. This emerging form of online shopping has introduced technical challenges, as products may be presented differently across various media domains. Therefore, a unified product representation is essential for achieving cross-domain product recognition to ensure an optimal user search experience and effective product recommendations. Despite the urgent industrial need for a unified cross-domain product representation, previous studies have predominantly focused only on product pages without taking into account short videos and live streams. To fill the gap in the rich-content e-commerce area, in this paper, we introduce a large-scale cRoss-dOmain Product Ecognition dataset, called ROPE. ROPE covers a wide range of product categories and contains over 180,000 products, corresponding to millions of short videos and live streams. It is the first dataset to cover product pages, short videos, and live streams simultaneously, providing the basis for establishing a unified product representation across different media domains. Furthermore, we propose a Cross-dOmain Product rEpresentation framework, namely COPE, which unifies product representations in different domains through multimodal learning including text and vision. Extensive experiments on downstream tasks demonstrate the effectiveness of COPE in learning a joint feature space for all product domains.
Authors: Danish Ebadulla, Rahul Raman, S. Natarajan, Hridhay Kiran Shetty, Ashish Harish Shenoy
Current research in zero-shot translation is plagued by several issues such as high compute requirements, increased training time and off target translations. Proposed remedies often come at the cost of additional data or compute requirements. Pivot based neural machine translation is preferred over a single-encoder model for most settings despite the increased training and evaluation time. In this work, we overcome the shortcomings of zero-shot translation by taking advantage of transliteration and linguistic similarity. We build a single encoder-decoder neural machine translation system for Dravidian-Dravidian multilingual translation and perform zero-shot translation. We compare the data vs zero-shot accuracy tradeoff and evaluate the performance of our vanilla method against the current state of the art pivot based method. We also test the theory that morphologically rich languages require large vocabularies by restricting the vocabulary using an optimal transport based technique. Our model manages to achieves scores within 3 BLEU of large-scale pivot-based models when it is trained on 50\% of the language directions.
Authors: Quan Tang, Chuanjian Liu, Fagui Liu, Yifan Liu, Jun Jiang, Bowen Zhang, Kai Han, Yunhe Wang
Aggregation of multi-stage features has been revealed to play a significant role in semantic segmentation. Unlike previous methods employing point-wise summation or concatenation for feature aggregation, this study proposes the Category Feature Transformer (CFT) that explores the flow of category embedding and transformation among multi-stage features through the prevalent multi-head attention mechanism. CFT learns unified feature embeddings for individual semantic categories from high-level features during each aggregation process and dynamically broadcasts them to high-resolution features. Integrating the proposed CFT into a typical feature pyramid structure exhibits superior performance over a broad range of backbone networks. We conduct extensive experiments on popular semantic segmentation benchmarks. Specifically, the proposed CFT obtains a compelling 55.1% mIoU with greatly reduced model parameters and computations on the challenging ADE20K dataset.
Authors: Alceu Bissoto, Catarina Barata, Eduardo Valle, Sandra Avila
Skin lesion analysis models are biased by artifacts placed during image acquisition, which influence model predictions despite carrying no clinical information. Solutions that address this problem by regularizing models to prevent learning those spurious features achieve only partial success, and existing test-time debiasing techniques are inappropriate for skin lesion analysis due to either making unrealistic assumptions on the distribution of test data or requiring laborious annotation from medical practitioners. We propose TTS (Test-Time Selection), a human-in-the-loop method that leverages positive (e.g., lesion area) and negative (e.g., artifacts) keypoints in test samples. TTS effectively steers models away from exploiting spurious artifact-related correlations without retraining, and with less annotation requirements. Our solution is robust to a varying availability of annotations, and different levels of bias. We showcase on the ISIC2019 dataset (for which we release a subset of annotated images) how our model could be deployed in the real-world for mitigating bias.
Authors: Edouard Yvinec, Arnaud Dapogny, Kevin Bailly
Deep neural network (DNN) deployment has been confined to larger hardware devices due to their expensive computational requirements. This challenge has recently reached another scale with the emergence of large language models (LLMs). In order to reduce both their memory footprint and latency, a promising technique is quantization. It consists in converting floating point representations to low bit-width fixed point representations, usually by assuming a uniform mapping onto a regular grid. This process, referred to in the literature as uniform quantization, may however be ill-suited as most DNN weights and activations follow a bell-shaped distribution. This is even worse on LLMs whose weight distributions are known to exhibit large, high impact, outlier values. In this work, we propose an improvement over the most commonly adopted way to tackle this limitation in deep learning models quantization, namely, non-uniform quantization. NUPES leverages automorphisms to preserve the scalar multiplications. Such transformations are derived from power functions. However, the optimization of the exponent parameter and weight values remains a challenging and novel problem which could not be solved with previous post training optimization techniques which only learn to round up or down weight values in order to preserve the predictive function. We circumvent this limitation with a new paradigm: learning new quantized weights over the entire quantized space. Similarly, we enable the optimization of the power exponent, i.e. the optimization of the quantization operator itself during training by alleviating all the numerical instabilities. The resulting predictive function is compatible with integer-only low-bit inference. We show the ability of the method to achieve state-of-the-art compression rates in both, data-free and data-driven configurations.
Authors: Shizhe Chen, Thomas Chabal, Ivan Laptev, Cordelia Schmid
Object goal navigation aims to navigate an agent to locations of a given object category in unseen environments. Classical methods explicitly build maps of environments and require extensive engineering while lacking semantic information for object-oriented exploration. On the other hand, end-to-end learning methods alleviate manual map design and predict actions using implicit representations. Such methods, however, lack an explicit notion of geometry and may have limited ability to encode navigation history. In this work, we propose an implicit spatial map for object goal navigation. Our implicit map is recursively updated with new observations at each step using a transformer. To encourage spatial reasoning, we introduce auxiliary tasks and train our model to reconstruct explicit maps as well as to predict visual features, semantic labels and actions. Our method significantly outperforms the state of the art on the challenging MP3D dataset and generalizes well to the HM3D dataset. We successfully deploy our model on a real robot and achieve encouraging object goal navigation results in real scenes using only a few real-world demonstrations. Code, trained models and videos are available at \url{https://www.di.ens.fr/willow/research/onav_rim/}.
Authors: Wencheng Han, Junbo Yin, Jianbing Shen
Monocular depth estimation is known as an ill-posed task in which objects in a 2D image usually do not contain sufficient information to predict their depth. Thus, it acts differently from other tasks (e.g., classification and segmentation) in many ways. In this paper, we find that self-supervised monocular depth estimation shows a direction sensitivity and environmental dependency in the feature representation. But the current backbones borrowed from other tasks pay less attention to handling different types of environmental information, limiting the overall depth accuracy. To bridge this gap, we propose a new Direction-aware Cumulative Convolution Network (DaCCN), which improves the depth feature representation in two aspects. First, we propose a direction-aware module, which can learn to adjust the feature extraction in each direction, facilitating the encoding of different types of information. Secondly, we design a new cumulative convolution to improve the efficiency for aggregating important environmental information. Experiments show that our method achieves significant improvements on three widely used benchmarks, KITTI, Cityscapes, and Make3D, setting a new state-of-the-art performance on the popular benchmarks with all three types of self-supervision.
Authors: Keqiang Fan, Xiaohao Cai, Mahesan Niranjan
Automated medical report generation has become increasingly important in medical analysis. It can produce computer-aided diagnosis descriptions and thus significantly alleviate the doctors' work. Inspired by the huge success of neural machine translation and image captioning, various deep learning methods have been proposed for medical report generation. However, due to the inherent properties of medical data, including data imbalance and the length and correlation between report sequences, the generated reports by existing methods may exhibit linguistic fluency but lack adequate clinical accuracy. In this work, we propose an image-to-indicator hierarchical transformer (IIHT) framework for medical report generation. It consists of three modules, i.e., a classifier module, an indicator expansion module and a generator module. The classifier module first extracts image features from the input medical images and produces disease-related indicators with their corresponding states. The disease-related indicators are subsequently utilised as input for the indicator expansion module, incorporating the "data-text-data" strategy. The transformer-based generator then leverages these extracted features along with image features as auxiliary information to generate final reports. Furthermore, the proposed IIHT method is feasible for radiologists to modify disease indicators in real-world scenarios and integrate the operations into the indicator expansion module for fluent and accurate medical report generation. Extensive experiments and comparisons with state-of-the-art methods under various evaluation metrics demonstrate the great performance of the proposed method.
Authors: Zezhong Lv, Bing Su, Ji-Rong Wen
Video moment localization aims to retrieve the target segment of an untrimmed video according to the natural language query. Weakly supervised methods gains attention recently, as the precise temporal location of the target segment is not always available. However, one of the greatest challenges encountered by the weakly supervised method is implied in the mismatch between the video and language induced by the coarse temporal annotations. To refine the vision-language alignment, recent works contrast the cross-modality similarities driven by reconstructing masked queries between positive and negative video proposals. However, the reconstruction may be influenced by the latent spurious correlation between the unmasked and the masked parts, which distorts the restoring process and further degrades the efficacy of contrastive learning since the masked words are not completely reconstructed from the cross-modality knowledge. In this paper, we discover and mitigate this spurious correlation through a novel proposed counterfactual cross-modality reasoning method. Specifically, we first formulate query reconstruction as an aggregated causal effect of cross-modality and query knowledge. Then by introducing counterfactual cross-modality knowledge into this aggregation, the spurious impact of the unmasked part contributing to the reconstruction is explicitly modeled. Finally, by suppressing the unimodal effect of masked query, we can rectify the reconstructions of video proposals to perform reasonable contrastive learning. Extensive experimental evaluations demonstrate the effectiveness of our proposed method. The code is available at \href{https://github.com/sLdZ0306/CCR}{https://github.com/sLdZ0306/CCR}.
Authors: Yanteng Zhang, Qizhi Teng, Xiaohai He, Tong Niu, Lipei Zhang, Yan Liu, Chao Ren
Structural MRI and PET imaging play an important role in the diagnosis of Alzheimer's disease (AD), showing the morphological changes and glucose metabolism changes in the brain respectively. The manifestations in the brain image of some cognitive impairment patients are relatively inconspicuous, for example, it still has difficulties in achieving accurate diagnosis through sMRI in clinical practice. With the emergence of deep learning, convolutional neural network (CNN) has become a valuable method in AD-aided diagnosis, but some CNN methods cannot effectively learn the features of brain image, making the diagnosis of AD still presents some challenges. In this work, we propose an end-to-end 3D CNN framework for AD diagnosis based on ResNet, which integrates multi-layer features obtained under the effect of the attention mechanism to better capture subtle differences in brain images. The attention maps showed our model can focus on key brain regions related to the disease diagnosis. Our method was verified in ablation experiments with two modality images on 792 subjects from the ADNI database, where AD diagnostic accuracies of 89.71% and 91.18% were achieved based on sMRI and PET respectively, and also outperformed some state-of-the-art methods.
Authors: Mainak Singha, Harsh Pal, Ankit Jha, Biplab Banerjee
Although deep learning models have shown impressive performance on supervised learning tasks, they often struggle to generalize well when the training (source) and test (target) domains differ. Unsupervised domain adaptation (DA) has emerged as a popular solution to this problem. However, current DA techniques rely on visual backbones, which may lack semantic richness. Despite the potential of large-scale vision-language foundation models like CLIP, their effectiveness for DA has yet to be fully explored. To address this gap, we introduce AD-CLIP, a domain-agnostic prompt learning strategy for CLIP that aims to solve the DA problem in the prompt space. We leverage the frozen vision backbone of CLIP to extract both image style (domain) and content information, which we apply to learn prompt tokens. Our prompts are designed to be domain-invariant and class-generalizable, by conditioning prompt learning on image style and content features simultaneously. We use standard supervised contrastive learning in the source domain, while proposing an entropy minimization strategy to align domains in the embedding space given the target domain data. We also consider a scenario where only target domain samples are available during testing, without any source domain data, and propose a cross-domain style mapping network to hallucinate domain-agnostic tokens. Our extensive experiments on three benchmark DA datasets demonstrate the effectiveness of AD-CLIP compared to existing literature.
Authors: Minhao Li, Zheng Qin, Zhirui Gao, Renjiao Yi, Chengyang Zhu, Kai Xu
The commonly adopted detect-then-match approach to registration finds difficulties in the cross-modality cases due to the incompatible keypoint detection and inconsistent feature description. We propose, 2D3D-MATR, a detection-free method for accurate and robust registration between images and point clouds. Our method adopts a coarse-to-fine pipeline where it first computes coarse correspondences between downsampled patches of the input image and the point cloud and then extends them to form dense correspondences between pixels and points within the patch region. The coarse-level patch matching is based on transformer which jointly learns global contextual constraints with self-attention and cross-modality correlations with cross-attention. To resolve the scale ambiguity in patch matching, we construct a multi-scale pyramid for each image patch and learn to find for each point patch the best matching image patch at a proper resolution level. Extensive experiments on two public benchmarks demonstrate that 2D3D-MATR outperforms the previous state-of-the-art P2-Net by around $20$ percentage points on inlier ratio and over $10$ points on registration recall. Our code and models are available at \url{https://github.com/minhaolee/2D3DMATR}.
Authors: Zhengzhi Lu, He Wang, Ziyi Chang, Guoan Yang, Hubert P. H. Shum
Recently, methods for skeleton-based human activity recognition have been shown to be vulnerable to adversarial attacks. However, these attack methods require either the full knowledge of the victim (i.e. white-box attacks), access to training data (i.e. transfer-based attacks) or frequent model queries (i.e. black-box attacks). All their requirements are highly restrictive, raising the question of how detrimental the vulnerability is. In this paper, we show that the vulnerability indeed exists. To this end, we consider a new attack task: the attacker has no access to the victim model or the training data or labels, where we coin the term hard no-box attack. Specifically, we first learn a motion manifold where we define an adversarial loss to compute a new gradient for the attack, named skeleton-motion-informed (SMI) gradient. Our gradient contains information of the motion dynamics, which is different from existing gradient-based attack methods that compute the loss gradient assuming each dimension in the data is independent. The SMI gradient can augment many gradient-based attack methods, leading to a new family of no-box attack methods. Extensive evaluation and comparison show that our method imposes a real threat to existing classifiers. They also show that the SMI gradient improves the transferability and imperceptibility of adversarial samples in both no-box and transfer-based black-box settings.
Authors: Zixuan Pan, Jianxu Chen, Yiyu Shi
Denoising diffusion probabilistic models have recently demonstrated state-of-the-art generative performance and been used as strong pixel-level representation learners. This paper decomposes the interrelation between the generative capability and representation learning ability inherent in diffusion models. We present masked diffusion model (MDM), a scalable self-supervised representation learner that substitutes the conventional additive Gaussian noise of traditional diffusion with a masking mechanism. Our proposed approach convincingly surpasses prior benchmarks, demonstrating remarkable advancements in both medical and natural image semantic segmentation tasks, particularly within the context of few-shot scenario.
Authors: Jiageng Zhu, Hanchen Xie, Jianhua Wu, Jiazhi Li, Mahyar Khayatkhoei, Mohamed E. Hussein, Wael AbdAlmageed
Discovering causal relations among semantic factors is an emergent topic in representation learning. Most causal representation learning (CRL) methods are fully supervised, which is impractical due to costly labeling. To resolve this restriction, weakly supervised CRL methods were introduced. To evaluate CRL performance, four existing datasets, Pendulum, Flow, CelebA(BEARD) and CelebA(SMILE), are utilized. However, existing CRL datasets are limited to simple graphs with few generative factors. Thus we propose two new datasets with a larger number of diverse generative factors and more sophisticated causal graphs. In addition, current real datasets, CelebA(BEARD) and CelebA(SMILE), the originally proposed causal graphs are not aligned with the dataset distributions. Thus, we propose modifications to them.
Authors: Yangyang Xu, Yibo Yang, Bernard Ghanemm, Lefei Zhang
CNNs and Transformers have their own advantages and both have been widely used for dense prediction in multi-task learning (MTL). Most of the current studies on MTL solely rely on CNN or Transformer. In this work, we present a novel MTL model by combining both merits of deformable CNN and query-based Transformer with shared gating for multi-task learning of dense prediction. This combination may offer a simple and efficient solution owing to its powerful and flexible task-specific learning and advantages of lower cost, less complexity and smaller parameters than the traditional MTL methods. We introduce deformable mixer Transformer with gating (DeMTG), a simple and effective encoder-decoder architecture up-to-date that incorporates the convolution and attention mechanism in a unified network for MTL. It is exquisitely designed to use advantages of each block, and provide deformable and comprehensive features for all tasks from local and global perspective. First, the deformable mixer encoder contains two types of operators: the channel-aware mixing operator leveraged to allow communication among different channels, and the spatial-aware deformable operator with deformable convolution applied to efficiently sample more informative spatial locations. Second, the task-aware gating transformer decoder is used to perform the task-specific predictions, in which task interaction block integrated with self-attention is applied to capture task interaction features, and the task query block integrated with gating attention is leveraged to select corresponding task-specific features. Further, the experiment results demonstrate that the proposed DeMTG uses fewer GFLOPs and significantly outperforms current Transformer-based and CNN-based competitive models on a variety of metrics on three dense prediction datasets. Our code and models are available at https://github.com/yangyangxu0/DeMTG.
Authors: Steffen Hagedorn, Marcel Hallgarten, Martin Stoll, Alexandru Condurache
Automated driving has the potential to revolutionize personal, public, and freight mobility. Besides the enormous challenge of perception, i.e. accurately perceiving the environment using available sensor data, automated driving comprises planning a safe, comfortable, and efficient motion trajectory. To promote safety and progress, many works rely on modules that predict the future motion of surrounding traffic. Modular automated driving systems commonly handle prediction and planning as sequential separate tasks. While this accounts for the influence of surrounding traffic on the ego-vehicle, it fails to anticipate the reactions of traffic participants to the ego-vehicle's behavior. Recent works suggest that integrating prediction and planning in an interdependent joint step is necessary to achieve safe, efficient, and comfortable driving. While various models implement such integrated systems, a comprehensive overview and theoretical understanding of different principles are lacking. We systematically review state-of-the-art deep learning-based prediction, planning, and integrated prediction and planning models. Different facets of the integration ranging from model architecture and model design to behavioral aspects are considered and related to each other. Moreover, we discuss the implications, strengths, and limitations of different integration methods. By pointing out research gaps, describing relevant future challenges, and highlighting trends in the research field, we identify promising directions for future research.
Authors: Guangkai Xu, Wei Yin, Hao Chen, Chunhua Shen, Kai Cheng, Feng Zhao
3D scene reconstruction is a long-standing vision task. Existing approaches can be categorized into geometry-based and learning-based methods. The former leverages multi-view geometry but can face catastrophic failures due to the reliance on accurate pixel correspondence across views. The latter was proffered to mitigate these issues by learning 2D or 3D representation directly. However, without a large-scale video or 3D training data, it can hardly generalize to diverse real-world scenarios due to the presence of tens of millions or even billions of optimization parameters in the deep network. Recently, robust monocular depth estimation models trained with large-scale datasets have been proven to possess weak 3D geometry prior, but they are insufficient for reconstruction due to the unknown camera parameters, the affine-invariant property, and inter-frame inconsistency. Here, we propose a novel test-time optimization approach that can transfer the robustness of affine-invariant depth models such as LeReS to challenging diverse scenes while ensuring inter-frame consistency, with only dozens of parameters to optimize per video frame. Specifically, our approach involves freezing the pre-trained affine-invariant depth model's depth predictions, rectifying them by optimizing the unknown scale-shift values with a geometric consistency alignment module, and employing the resulting scale-consistent depth maps to robustly obtain camera poses and achieve dense scene reconstruction, even in low-texture regions. Experiments show that our method achieves state-of-the-art cross-dataset reconstruction on five zero-shot testing datasets.
Authors: Bencheng Liao, Shaoyu Chen, Yunchi Zhang, Bo Jiang, Qian Zhang, Wenyu Liu, Chang Huang, Xinggang Wang
High-definition (HD) map provides abundant and precise static environmental information of the driving scene, serving as a fundamental and indispensable component for planning in autonomous driving system. In this paper, we present \textbf{Map} \textbf{TR}ansformer, an end-to-end framework for online vectorized HD map construction. We propose a unified permutation-equivalent modeling approach, \ie, modeling map element as a point set with a group of equivalent permutations, which accurately describes the shape of map element and stabilizes the learning process. We design a hierarchical query embedding scheme to flexibly encode structured map information and perform hierarchical bipartite matching for map element learning. To speed up convergence, we further introduce auxiliary one-to-many matching and dense supervision. The proposed method well copes with various map elements with arbitrary shapes. It runs at real-time inference speed and achieves state-of-the-art performance on both nuScenes and Argoverse2 datasets. Abundant qualitative results show stable and robust map construction quality in complex and various driving scenes. Code and more demos are available at \url{https://github.com/hustvl/MapTR} for facilitating further studies and applications.
Authors: Alaa Maalouf, Ninad Jadhav, Krishna Murthy Jatavallabhula, Makram Chahine, Daniel M.Vogt, Robert J. Wood, Antonio Torralba, Daniela Rus
Tracking and following objects of interest is critical to several robotics use cases, ranging from industrial automation to logistics and warehousing, to healthcare and security. In this paper, we present a robotic system to detect, track, and follow any object in real-time. Our approach, dubbed ``follow anything'' (FAn), is an open-vocabulary and multimodal model -- it is not restricted to concepts seen at training time and can be applied to novel classes at inference time using text, images, or click queries. Leveraging rich visual descriptors from large-scale pre-trained models (foundation models), FAn can detect and segment objects by matching multimodal queries (text, images, clicks) against an input image sequence. These detected and segmented objects are tracked across image frames, all while accounting for occlusion and object re-emergence. We demonstrate FAn on a real-world robotic system (a micro aerial vehicle) and report its ability to seamlessly follow the objects of interest in a real-time control loop. FAn can be deployed on a laptop with a lightweight (6-8 GB) graphics card, achieving a throughput of 6-20 frames per second. To enable rapid adoption, deployment, and extensibility, we open-source all our code on our project webpage at https://github.com/alaamaalouf/FollowAnything . We also encourage the reader the watch our 5-minutes explainer video in this https://www.youtube.com/watch?v=6Mgt3EPytrw .
Authors: Michael Fischer, Tobias Ritschel
Gradient-based optimization is now ubiquitous across graphics, but unfortunately can not be applied to problems with undefined or zero gradients. To circumvent this issue, the loss function can be manually replaced by a "surrogate" that has similar minima but is differentiable. Our proposed framework, ZeroGrads, automates this process by learning a neural approximation of the objective function, the surrogate, which in turn can be used to differentiate through arbitrary black-box graphics pipelines. We train the surrogate on an actively smoothed version of the objective and encourage locality, focusing the surrogate's capacity on what matters at the current training episode. The fitting is performed online, alongside the parameter optimization, and self-supervised, without pre-computed data or pre-trained models. As sampling the objective is expensive (it requires a full rendering or simulator run), we devise an efficient sampling scheme that allows for tractable run-times and competitive performance at little overhead. We demonstrate optimizing diverse non-convex, non-differentiable black-box problems in graphics, such as visibility in rendering, discrete parameter spaces in procedural modelling or optimal control in physics-driven animation. In contrast to more traditional algorithms, our approach scales well to higher dimensions, which we demonstrate on problems with up to 35k interlinked variables.
Authors: Yun-Chun Chen, Vladimir G. Kim, Noam Aigerman, Alec Jacobson
The recent proliferation of 3D content that can be consumed on hand-held devices necessitates efficient tools for transmitting large geometric data, e.g., 3D meshes, over the Internet. Detailed high-resolution assets can pose a challenge to storage as well as transmission bandwidth, and level-of-detail techniques are often used to transmit an asset using an appropriate bandwidth budget. It is especially desirable for these methods to transmit data progressively, improving the quality of the geometry with more data. Our key insight is that the geometric details of 3D meshes often exhibit similar local patterns even across different shapes, and thus can be effectively represented with a shared learned generative space. We learn this space using a subdivision-based encoder-decoder architecture trained in advance on a large collection of surfaces. We further observe that additional residual features can be transmitted progressively between intermediate levels of subdivision that enable the client to control the tradeoff between bandwidth cost and quality of reconstruction, providing a neural progressive mesh representation. We evaluate our method on a diverse set of complex 3D shapes and demonstrate that it outperforms baselines in terms of compression ratio and reconstruction quality.
Authors: Wentao Hu, Jia Zheng, Zixin Zhang, Xiaojun Yuan, Jian Yin, Zihan Zhou
In this paper, we develop a new method to automatically convert 2D line drawings from three orthographic views into 3D CAD models. Existing methods for this problem reconstruct 3D models by back-projecting the 2D observations into 3D space while maintaining explicit correspondence between the input and output. Such methods are sensitive to errors and noises in the input, thus often fail in practice where the input drawings created by human designers are imperfect. To overcome this difficulty, we leverage the attention mechanism in a Transformer-based sequence generation model to learn flexible mappings between the input and output. Further, we design shape programs which are suitable for generating the objects of interest to boost the reconstruction accuracy and facilitate CAD modeling applications. Experiments on a new benchmark dataset show that our method significantly outperforms existing ones when the inputs are noisy or incomplete.
Authors: Iaroslav Koshelev, Stamatios Lefkimmiatis
In this work we present a novel optimization strategy for image reconstruction tasks under analysis-based image regularization, which promotes sparse and/or low-rank solutions in some learned transform domain. We parameterize such regularizers using potential functions that correspond to weighted extensions of the $\ell_p^p$-vector and $\mathcal{S}_p^p$ Schatten-matrix quasi-norms with $0 < p \le 1$. Our proposed minimization strategy extends the Iteratively Reweighted Least Squares (IRLS) method, typically used for synthesis-based $\ell_p$ and $\mathcal{S}_p$ norm and analysis-based $\ell_1$ and nuclear norm regularization. We prove that under mild conditions our minimization algorithm converges linearly to a stationary point, and we provide an upper bound for its convergence rate. Further, to select the parameters of the regularizers that deliver the best results for the problem at hand, we propose to learn them from training data by formulating the supervised learning process as a stochastic bilevel optimization problem. We show that thanks to the convergence guarantees of our proposed minimization strategy, such optimization can be successfully performed with a memory-efficient implicit back-propagation scheme. We implement our learned IRLS variants as recurrent networks and assess their performance on the challenging image reconstruction tasks of non-blind deblurring, super-resolution and demosaicking. The comparisons against other existing learned reconstruction approaches demonstrate that our overall method is very competitive and in many cases outperforms existing unrolled networks, whose number of parameters is orders of magnitude higher than in our case.
Authors: Jiong Wu, Xiaoying Tang
In this study, we proposed and validated a multi-atlas guided 3D fully convolutional network (FCN) ensemble model (M-FCN) for segmenting brain regions of interest (ROIs) from structural magnetic resonance images (MRIs). One major limitation of existing state-of-the-art 3D FCN segmentation models is that they often apply image patches of fixed size throughout training and testing, which may miss some complex tissue appearance patterns of different brain ROIs. To address this limitation, we trained a 3D FCN model for each ROI using patches of adaptive size and embedded outputs of the convolutional layers in the deconvolutional layers to further capture the local and global context patterns. In addition, with an introduction of multi-atlas based guidance in M-FCN, our segmentation was generated by combining the information of images and labels, which is highly robust. To reduce over-fitting of the FCN model on the training data, we adopted an ensemble strategy in the learning procedure. Evaluation was performed on two brain MRI datasets, aiming respectively at segmenting 14 subcortical and ventricular structures and 54 brain ROIs. The segmentation results of the proposed method were compared with those of a state-of-the-art multi-atlas based segmentation method and an existing 3D FCN segmentation model. Our results suggested that the proposed method had a superior segmentation performance.
Authors: Longtian Qiu, Renrui Zhang, Ziyu Guo, Ziyao Zeng, Zilu Guo, Yafeng Li, Guangnan Zhang
Contrastive Language-Image Pre-training (CLIP) has drawn increasing attention recently for its transferable visual representation learning. However, due to the semantic gap within datasets, CLIP's pre-trained image-text alignment becomes sub-optimal on downstream tasks, which severely harms its transferring performance. To better adapt the cross-modality embedding space, we propose to enhance CLIP via Visual-guided Texts, named VT-CLIP. Specifically, we guide textual features of different categories to adaptively explore informative regions on the image and aggregate visual features by attention mechanisms. In this way, the texts become visual-guided, namely, more semantically correlated with downstream images, which greatly benefits the category-wise matching process. In few-shot settings, we evaluate our VT-CLIP on 11 well-known classification datasets to demonstrate its effectiveness.
Authors: Minh-Quan Le, Trung-Nghia Le, Tam V. Nguyen, Isao Echizen, Minh-Triet Tran
Recent years have witnessed great advances in object segmentation research. In addition to generic objects, aquatic animals have attracted research attention. Deep learning-based methods are widely used for aquatic animal segmentation and have achieved promising performance. However, there is a lack of challenging datasets for benchmarking. In this work, we build a new dataset dubbed Aquatic Animal Species. We also devise a novel GUided mixup augmeNtatioN and multi-modEl fusion for aquatic animaL segmentation (GUNNEL) that leverages the advantages of multiple segmentation models to effectively segment aquatic animals and improves the training performance by synthesizing hard samples. Extensive experiments demonstrated the superiority of our proposed framework over existing state-of-the-art instance segmentation methods. The code is available at https://github.com/lmquan2000/mask-mixup. The dataset is available at https://doi.org/10.5281/zenodo.8208877 .
Authors: Xiaokang Chen, Mingyu Ding, Xiaodi Wang, Ying Xin, Shentong Mo, Yunhao Wang, Shumin Han, Ping Luo, Gang Zeng, Jingdong Wang
We present a novel masked image modeling (MIM) approach, context autoencoder (CAE), for self-supervised representation pretraining. We pretrain an encoder by making predictions in the encoded representation space. The pretraining tasks include two tasks: masked representation prediction - predict the representations for the masked patches, and masked patch reconstruction - reconstruct the masked patches. The network is an encoder-regressor-decoder architecture: the encoder takes the visible patches as input; the regressor predicts the representations of the masked patches, which are expected to be aligned with the representations computed from the encoder, using the representations of visible patches and the positions of visible and masked patches; the decoder reconstructs the masked patches from the predicted encoded representations. The CAE design encourages the separation of learning the encoder (representation) from completing the pertaining tasks: masked representation prediction and masked patch reconstruction tasks, and making predictions in the encoded representation space empirically shows the benefit to representation learning. We demonstrate the effectiveness of our CAE through superior transfer performance in downstream tasks: semantic segmentation, object detection and instance segmentation, and classification. The code will be available at https://github.com/Atten4Vis/CAE.
Authors: Nathan Michlo, Richard Klein, Steven James
Learning disentangled representations with variational autoencoders (VAEs) is often attributed to the regularisation component of the loss. In this work, we highlight the interaction between data and the reconstruction term of the loss as the main contributor to disentanglement in VAEs. We show that standard benchmark datasets have unintended correlations between their subjective ground-truth factors and perceived axes in the data according to typical VAE reconstruction losses. Our work exploits this relationship to provide a theory for what constitutes an adversarial dataset under a given reconstruction loss. We verify this by constructing an example dataset that prevents disentanglement in state-of-the-art frameworks while maintaining human-intuitive ground-truth factors. Finally, we re-enable disentanglement by designing an example reconstruction loss that is once again able to perceive the ground-truth factors. Our findings demonstrate the subjective nature of disentanglement and the importance of considering the interaction between the ground-truth factors, data and notably, the reconstruction loss, which is under-recognised in the literature.
Authors: Jiashuo Yu, Junfu Pu, Ying Cheng, Rui Feng, Ying Shan
Although audio-visual representation has been proved to be applicable in many downstream tasks, the representation of dancing videos, which is more specific and always accompanied by music with complex auditory contents, remains challenging and uninvestigated. Considering the intrinsic alignment between the cadent movement of dancer and music rhythm, we introduce MuDaR, a novel Music-Dance Representation learning framework to perform the synchronization of music and dance rhythms both in explicit and implicit ways. Specifically, we derive the dance rhythms based on visual appearance and motion cues inspired by the music rhythm analysis. Then the visual rhythms are temporally aligned with the music counterparts, which are extracted by the amplitude of sound intensity. Meanwhile, we exploit the implicit coherence of rhythms implied in audio and visual streams by contrastive learning. The model learns the joint embedding by predicting the temporal consistency between audio-visual pairs. The music-dance representation, together with the capability of detecting audio and visual rhythms, can further be applied to three downstream tasks: (a) dance classification, (b) music-dance retrieval, and (c) music-dance retargeting. Extensive experiments demonstrate that our proposed framework outperforms other self-supervised methods by a large margin.
Authors: Rajitha de Silva, Grzegorz Cielniak, Gang Wang, Junfeng Gao
Autonomous navigation in agricultural environments is challenged by varying field conditions that arise in arable fields. State-of-the-art solutions for autonomous navigation in such environments require expensive hardware such as RTK-GNSS. This paper presents a robust crop row detection algorithm that withstands such field variations using inexpensive cameras. Existing datasets for crop row detection does not represent all the possible field variations. A dataset of sugar beet images was created representing 11 field variations comprised of multiple grow stages, light levels, varying weed densities, curved crop rows and discontinuous crop rows. The proposed pipeline segments the crop rows using a deep learning-based method and employs the predicted segmentation mask for extraction of the central crop using a novel central crop row selection algorithm. The novel crop row detection algorithm was tested for crop row detection performance and the capability of visual servoing along a crop row. The visual servoing-based navigation was tested on a realistic simulation scenario with the real ground and plant textures. Our algorithm demonstrated robust vision-based crop row detection in challenging field conditions outperforming the baseline.
Authors: Jiangmeng Li, Wenwen Qiang, Yanan Zhang, Wenyi Mo, Changwen Zheng, Bing Su, Hui Xiong
As a successful approach to self-supervised learning, contrastive learning aims to learn invariant information shared among distortions of the input sample. While contrastive learning has yielded continuous advancements in sampling strategy and architecture design, it still remains two persistent defects: the interference of task-irrelevant information and sample inefficiency, which are related to the recurring existence of trivial constant solutions. From the perspective of dimensional analysis, we find out that the dimensional redundancy and dimensional confounder are the intrinsic issues behind the phenomena, and provide experimental evidence to support our viewpoint. We further propose a simple yet effective approach MetaMask, short for the dimensional Mask learned by Meta-learning, to learn representations against dimensional redundancy and confounder. MetaMask adopts the redundancy-reduction technique to tackle the dimensional redundancy issue and innovatively introduces a dimensional mask to reduce the gradient effects of specific dimensions containing the confounder, which is trained by employing a meta-learning paradigm with the objective of improving the performance of masked representations on a typical self-supervised task. We provide solid theoretical analyses to prove MetaMask can obtain tighter risk bounds for downstream classification compared to typical contrastive methods. Empirically, our method achieves state-of-the-art performance on various benchmarks.
Authors: Eddy Zhou, Alex Zhuang, Alikasim Budhwani, Rowan Dempster, Quanquan Li, Mohammad Al-Sharman, Derek Rayside, William Melek
When applied to autonomous vehicle (AV) settings, action recognition can enhance an environment model's situational awareness. This is especially prevalent in scenarios where traditional geometric descriptions and heuristics in AVs are insufficient. However, action recognition has traditionally been studied for humans, and its limited adaptability to noisy, un-clipped, un-pampered, raw RGB data has limited its application in other fields. To push for the advancement and adoption of action recognition into AVs, this work proposes a novel two-stage action recognition system, termed RALACs. RALACs formulates the problem of action recognition for road scenes, and bridges the gap between it and the established field of human action recognition. This work shows how attention layers can be useful for encoding the relations across agents, and stresses how such a scheme can be class-agnostic. Furthermore, to address the dynamic nature of agents on the road, RALACs constructs a novel approach to adapting Region of Interest (ROI) Alignment to agent tracks for downstream action classification. Finally, our scheme also considers the problem of active agent detection, and utilizes a novel application of fusing optical flow maps to discern relevant agents in a road scene. We show that our proposed scheme can outperform the baseline on the ICCV2021 Road Challenge dataset and by deploying it on a real vehicle platform, we provide preliminary insight to the usefulness of action recognition in decision making.
Authors: Blerta Lindqvist
This paper uses symmetry to make Convolutional Neural Network classifiers (CNNs) robust against adversarial perturbation attacks. Such attacks add perturbation to original images to generate adversarial images that fool classifiers such as road sign classifiers of autonomous vehicles. Although symmetry is a pervasive aspect of the natural world, CNNs are unable to handle symmetry well. For example, a CNN can classify an image differently from its mirror image. For an adversarial image that misclassifies with a wrong label $l_w$, CNN inability to handle symmetry means that a symmetric adversarial image can classify differently from the wrong label $l_w$. Further than that, we find that the classification of a symmetric adversarial image reverts to the correct label. To classify an image when adversaries are unaware of the defense, we apply symmetry to the image and use the classification label of the symmetric image. To classify an image when adversaries are aware of the defense, we use mirror symmetry and pixel inversion symmetry to form a symmetry group. We apply all the group symmetries to the image and decide on the output label based on the agreement of any two of the classification labels of the symmetry images. Adaptive attacks fail because they need to rely on loss functions that use conflicting CNN output values for symmetric images. Without attack knowledge, the proposed symmetry defense succeeds against both gradient-based and random-search attacks, with up to near-default accuracies for ImageNet. The defense even improves the classification accuracy of original images.
Authors: Kyle Min
This report describes our approach for the Audio-Visual Diarization (AVD) task of the Ego4D Challenge 2022. Specifically, we present multiple technical improvements over the official baselines. First, we improve the detection performance of the camera wearer's voice activity by modifying the training scheme of its model. Second, we discover that an off-the-shelf voice activity detection model can effectively remove false positives when it is applied solely to the camera wearer's voice activities. Lastly, we show that better active speaker detection leads to a better AVD outcome. Our final method obtains 65.9% DER on the test set of Ego4D, which significantly outperforms all the baselines. Our submission achieved 1st place in the Ego4D Challenge 2022.
Authors: Ryuichiro Hataya, Han Bao, Hiromi Arai
Recently proposed large-scale text-to-image generative models such as DALL$\cdot$E 2, Midjourney, and StableDiffusion can generate high-quality and realistic images from users' prompts. Not limited to the research community, ordinary Internet users enjoy these generative models, and consequently, a tremendous amount of generated images have been shared on the Internet. Meanwhile, today's success of deep learning in the computer vision field owes a lot to images collected from the Internet. These trends lead us to a research question: "\textbf{will such generated images impact the quality of future datasets and the performance of computer vision models positively or negatively?}" This paper empirically answers this question by simulating contamination. Namely, we generate ImageNet-scale and COCO-scale datasets using a state-of-the-art generative model and evaluate models trained with "contaminated" datasets on various tasks, including image classification and image generation. Throughout experiments, we conclude that generated images negatively affect downstream performance, while the significance depends on tasks and the amount of generated images. The generated datasets and the codes for experiments will be publicly released for future research. Generated datasets and source codes are available from \url{https://github.com/moskomule/dataset-contamination}.
Authors: Zhiyang Dou, Qingxuan Wu, Cheng Lin, Zeyu Cao, Qiangqiang Wu, Weilin Wan, Taku Komura, Wenping Wang
In this paper, we introduce a set of simple yet effective TOken REduction (TORE) strategies for Transformer-based Human Mesh Recovery from monocular images. Current SOTA performance is achieved by Transformer-based structures. However, they suffer from high model complexity and computation cost caused by redundant tokens. We propose token reduction strategies based on two important aspects, i.e., the 3D geometry structure and 2D image feature, where we hierarchically recover the mesh geometry with priors from body structure and conduct token clustering to pass fewer but more discriminative image feature tokens to the Transformer. Our method massively reduces the number of tokens involved in high-complexity interactions in the Transformer. This leads to a significantly reduced computational cost while still achieving competitive or even higher accuracy in shape recovery. Extensive experiments across a wide range of benchmarks validate the superior effectiveness of the proposed method. We further demonstrate the generalizability of our method on hand mesh recovery. Visit our project page at https://frank-zy-dou.github.io/projects/Tore/index.html.
Authors: Ziyi Li, Qinye Zhou, Xiaoyun Zhang, Ya Zhang, Yanfeng Wang, Weidi Xie
The goal of this paper is to extract the visual-language correspondence from a pre-trained text-to-image diffusion model, in the form of segmentation map, i.e., simultaneously generating images and segmentation masks for the corresponding visual entities described in the text prompt. We make the following contributions: (i) we pair the existing Stable Diffusion model with a novel grounding module, that can be trained to align the visual and textual embedding space of the diffusion model with only a small number of object categories; (ii) we establish an automatic pipeline for constructing a dataset, that consists of {image, segmentation mask, text prompt} triplets, to train the proposed grounding module; (iii) we evaluate the performance of open-vocabulary grounding on images generated from the text-to-image diffusion model and show that the module can well segment the objects of categories beyond seen ones at training time; (iv) we adopt the augmented diffusion model to build a synthetic semantic segmentation dataset, and show that, training a standard segmentation model on such dataset demonstrates competitive performance on the zero-shot segmentation(ZS3) benchmark, which opens up new opportunities for adopting the powerful diffusion model for discriminative tasks.
Authors: George Yiasemis, Clara I. Sánchez, Jan-Jakob Sonke, Jonas Teuwen
Acquiring fully-sampled MRI $k$-space data is time-consuming, and collecting accelerated data can reduce the acquisition time. Employing 2D Cartesian-rectilinear subsampling schemes is a conventional approach for accelerated acquisitions; however, this often results in imprecise reconstructions, even with the use of Deep Learning (DL), especially at high acceleration factors. Non-rectilinear or non-Cartesian trajectories can be implemented in MRI scanners as alternative subsampling options. This work investigates the impact of the $k$-space subsampling scheme on the quality of reconstructed accelerated MRI measurements produced by trained DL models. The Recurrent Variational Network (RecurrentVarNet) was used as the DL-based MRI-reconstruction architecture. Cartesian, fully-sampled multi-coil $k$-space measurements from three datasets were retrospectively subsampled with different accelerations using eight distinct subsampling schemes: four Cartesian-rectilinear, two Cartesian non-rectilinear, and two non-Cartesian. Experiments were conducted in two frameworks: scheme-specific, where a distinct model was trained and evaluated for each dataset-subsampling scheme pair, and multi-scheme, where for each dataset a single model was trained on data randomly subsampled by any of the eight schemes and evaluated on data subsampled by all schemes. In both frameworks, RecurrentVarNets trained and evaluated on non-rectilinearly subsampled data demonstrated superior performance, particularly for high accelerations. In the multi-scheme setting, reconstruction performance on rectilinearly subsampled data improved when compared to the scheme-specific experiments. Our findings demonstrate the potential for using DL-based methods, trained on non-rectilinearly subsampled measurements, to optimize scan time and image quality.
Authors: David Bruggemann, Christos Sakaridis, Tim Brödermann, Luc Van Gool
Standard unsupervised domain adaptation methods adapt models from a source to a target domain using labeled source data and unlabeled target data jointly. In model adaptation, on the other hand, access to the labeled source data is prohibited, i.e., only the source-trained model and unlabeled target data are available. We investigate normal-to-adverse condition model adaptation for semantic segmentation, whereby image-level correspondences are available in the target domain. The target set consists of unlabeled pairs of adverse- and normal-condition street images taken at GPS-matched locations. Our method -- CMA -- leverages such image pairs to learn condition-invariant features via contrastive learning. In particular, CMA encourages features in the embedding space to be grouped according to their condition-invariant semantic content and not according to the condition under which respective inputs are captured. To obtain accurate cross-domain semantic correspondences, we warp the normal image to the viewpoint of the adverse image and leverage warp-confidence scores to create robust, aggregated features. With this approach, we achieve state-of-the-art semantic segmentation performance for model adaptation on several normal-to-adverse adaptation benchmarks, such as ACDC and Dark Zurich. We also evaluate CMA on a newly procured adverse-condition generalization benchmark and report favorable results compared to standard unsupervised domain adaptation methods, despite the comparative handicap of CMA due to source data inaccessibility. Code is available at https://github.com/brdav/cma.
Authors: Siquan Huang, Yijiang Li, Chong Chen, Leyu Shi, Ying Gao
The decentralized and privacy-preserving nature of federated learning (FL) makes it vulnerable to backdoor attacks aiming to manipulate the behavior of the resulting model on specific adversary-chosen inputs. However, most existing defenses based on statistical differences take effect only against specific attacks, especially when the malicious gradients are similar to benign ones or the data are highly non-independent and identically distributed (non-IID). In this paper, we revisit the distance-based defense methods and discover that i) Euclidean distance becomes meaningless in high dimensions and ii) malicious gradients with diverse characteristics cannot be identified by a single metric. To this end, we present a simple yet effective defense strategy with multi-metrics and dynamic weighting to identify backdoors adaptively. Furthermore, our novel defense has no reliance on predefined assumptions over attack settings or data distributions and little impact on benign performance. To evaluate the effectiveness of our approach, we conduct comprehensive experiments on different datasets under various attack settings, where our method achieves the best defensive performance. For instance, we achieve the lowest backdoor accuracy of 3.06% under the difficult Edge-case PGD, showing significant superiority over previous defenses. The results also demonstrate that our method can be well-adapted to a wide range of non-IID degrees without sacrificing the benign performance.
Authors: Jiaqi Chen, Jiachen Lu, Xiatian Zhu, Li Zhang
We present Generative Semantic Segmentation (GSS), a generative learning approach for semantic segmentation. Uniquely, we cast semantic segmentation as an image-conditioned mask generation problem. This is achieved by replacing the conventional per-pixel discriminative learning with a latent prior learning process. Specifically, we model the variational posterior distribution of latent variables given the segmentation mask. To that end, the segmentation mask is expressed with a special type of image (dubbed as maskige). This posterior distribution allows to generate segmentation masks unconditionally. To achieve semantic segmentation on a given image, we further introduce a conditioning network. It is optimized by minimizing the divergence between the posterior distribution of maskige (i.e., segmentation masks) and the latent prior distribution of input training images. Extensive experiments on standard benchmarks show that our GSS can perform competitively to prior art alternatives in the standard semantic segmentation setting, whilst achieving a new state of the art in the more challenging cross-domain setting.
Authors: Jiuming Liu, Guangming Wang, Zhe Liu, Chaokang Jiang, Marc Pollefeys, Hesheng Wang
Although point cloud registration has achieved remarkable advances in object-level and indoor scenes, large-scale registration methods are rarely explored. Challenges mainly arise from the huge point number, complex distribution, and outliers of outdoor LiDAR scans. In addition, most existing registration works generally adopt a two-stage paradigm: They first find correspondences by extracting discriminative local features and then leverage estimators (eg. RANSAC) to filter outliers, which are highly dependent on well-designed descriptors and post-processing choices. To address these problems, we propose an end-to-end transformer network (RegFormer) for large-scale point cloud alignment without any further post-processing. Specifically, a projection-aware hierarchical transformer is proposed to capture long-range dependencies and filter outliers by extracting point features globally. Our transformer has linear complexity, which guarantees high efficiency even for large-scale scenes. Furthermore, to effectively reduce mismatches, a bijective association transformer is designed for regressing the initial transformation. Extensive experiments on KITTI and NuScenes datasets demonstrate that our RegFormer achieves competitive performance in terms of both accuracy and efficiency.
Authors: Nicola Franco, Daniel Korth, Jeanette Miriam Lorenz, Karsten Roscher, Stephan Guennemann
As the use of machine learning continues to expand, the importance of ensuring its safety cannot be overstated. A key concern in this regard is the ability to identify whether a given sample is from the training distribution, or is an "Out-Of-Distribution" (OOD) sample. In addition, adversaries can manipulate OOD samples in ways that lead a classifier to make a confident prediction. In this study, we present a novel approach for certifying the robustness of OOD detection within a $\ell_2$-norm around the input, regardless of network architecture and without the need for specific components or additional training. Further, we improve current techniques for detecting adversarial attacks on OOD samples, while providing high levels of certified and adversarial robustness on in-distribution samples. The average of all OOD detection metrics on CIFAR10/100 shows an increase of $\sim 13 \% / 5\%$ relative to previous approaches.
Authors: Lin Zhao, Mingxi Zhou, Brice Loose
Robotic underwater systems, e.g., Autonomous Underwater Vehicles (AUVs) and Remotely Operated Vehicles (ROVs), are promising tools for collecting biogeochemical data at the ice-water interface for scientific advancements. However, state estimation, i.e., localization, is a well-known problem for robotic systems, especially, for the ones that travel underwater. In this paper, we present a tightly-coupled multi-sensors fusion framework to increase localization accuracy that is robust to sensor failure. Visual images, Doppler Velocity Log (DVL), Inertial Measurement Unit (IMU) and Pressure sensor are integrated into the state-of-art Multi-State Constraint Kalman Filter (MSCKF) for state estimation. Besides that a new keyframe-based state clone mechanism and a new DVL-aided feature enhancement are presented to further improve the localization performance. The proposed method is validated with a data set collected in the field under frozen ice, and the result is compared with 6 other different sensor fusion setups. Overall, the result with the keyframe enabled and DVL-aided feature enhancement yields the best performance with a Root-mean-square error of less than 2 m compared to the ground truth path with a total traveling distance of about 200 m.
Authors: Yuanhong Chen, Yuyuan Liu, Hu Wang, Fengbei Liu, Chong Wang, Gustavo Carneiro
Audio-visual segmentation (AVS) is a complex task that involves accurately segmenting the corresponding sounding object based on audio-visual queries. Successful audio-visual learning requires two essential components: 1) an unbiased dataset with high-quality pixel-level multi-class labels, and 2) a model capable of effectively linking audio information with its corresponding visual object. However, these two requirements are only partially addressed by current methods, with training sets containing biased audio-visual data, and models that generalise poorly beyond this biased training set. In this work, we propose a new strategy to build cost-effective and relatively unbiased audio-visual semantic segmentation benchmarks. Our strategy, called Visual Post-production (VPO), explores the observation that it is not necessary to have explicit audio-visual pairs extracted from single video sources to build such benchmarks. We also refine the previously proposed AVSBench to transform it into the audio-visual semantic segmentation benchmark AVSBench-Single+. Furthermore, this paper introduces a new pixel-wise audio-visual contrastive learning method to enable a better generalisation of the model beyond the training set. We verify the validity of the VPO strategy by showing that state-of-the-art (SOTA) models trained with datasets built by matching audio and visual data from different sources or with datasets containing audio and visual data from the same video source produce almost the same accuracy. Then, using the proposed VPO benchmarks and AVSBench-Single+, we show that our method produces more accurate audio-visual semantic segmentation than SOTA models. Code and dataset will be available.
Authors: Ahmet Burak Yildirim, Vedat Baday, Erkut Erdem, Aykut Erdem, Aysegul Dundar
Image inpainting task refers to erasing unwanted pixels from images and filling them in a semantically consistent and realistic way. Traditionally, the pixels that are wished to be erased are defined with binary masks. From the application point of view, a user needs to generate the masks for the objects they would like to remove which can be time-consuming and prone to errors. In this work, we are interested in an image inpainting algorithm that estimates which object to be removed based on natural language input and removes it, simultaneously. For this purpose, first, we construct a dataset named GQA-Inpaint for this task. Second, we present a novel inpainting framework, Inst-Inpaint, that can remove objects from images based on the instructions given as text prompts. We set various GAN and diffusion-based baselines and run experiments on synthetic and real image datasets. We compare methods with different evaluation metrics that measure the quality and accuracy of the models and show significant quantitative and qualitative improvements.
Authors: Jie Chen, Hua Mao, Wai Lok Woo, Xi Peng
Multiview clustering (MVC) aims to reveal the underlying structure of multiview data by categorizing data samples into clusters. Deep learning-based methods exhibit strong feature learning capabilities on large-scale datasets. For most existing deep MVC methods, exploring the invariant representations of multiple views is still an intractable problem. In this paper, we propose a cross-view contrastive learning (CVCL) method that learns view-invariant representations and produces clustering results by contrasting the cluster assignments among multiple views. Specifically, we first employ deep autoencoders to extract view-dependent features in the pretraining stage. Then, a cluster-level CVCL strategy is presented to explore consistent semantic label information among the multiple views in the fine-tuning stage. Thus, the proposed CVCL method is able to produce more discriminative cluster assignments by virtue of this learning strategy. Moreover, we provide a theoretical analysis of soft cluster assignment alignment. Extensive experimental results obtained on several datasets demonstrate that the proposed CVCL method outperforms several state-of-the-art approaches.
Authors: Ruixiang Jiang, Lingbo Liu, Changwen Chen
Recent advances in visual-language models have shown remarkable zero-shot text-image matching ability that is transferable to downstream tasks such as object detection and segmentation. Adapting these models for object counting, however, remains a formidable challenge. In this study, we first investigate transferring vision-language models (VLMs) for class-agnostic object counting. Specifically, we propose CLIP-Count, the first end-to-end pipeline that estimates density maps for open-vocabulary objects with text guidance in a zero-shot manner. To align the text embedding with dense visual features, we introduce a patch-text contrastive loss that guides the model to learn informative patch-level visual representations for dense prediction. Moreover, we design a hierarchical patch-text interaction module to propagate semantic information across different resolution levels of visual features. Benefiting from the full exploitation of the rich image-text alignment knowledge of pretrained VLMs, our method effectively generates high-quality density maps for objects-of-interest. Extensive experiments on FSC-147, CARPK, and ShanghaiTech crowd counting datasets demonstrate state-of-the-art accuracy and generalizability of the proposed method. Code is available: https://github.com/songrise/CLIP-Count.
Authors: Kyle Min
This report introduces our novel method named STHG for the Audio-Visual Diarization task of the Ego4D Challenge 2023. Our key innovation is that we model all the speakers in a video using a single, unified heterogeneous graph learning framework. Unlike previous approaches that require a separate component solely for the camera wearer, STHG can jointly detect the speech activities of all people including the camera wearer. Our final method obtains 61.1% DER on the test set of Ego4D, which significantly outperforms all the baselines as well as last year's winner. Our submission achieved 1st place in the Ego4D Challenge 2023. We additionally demonstrate that applying the off-the-shelf speech recognition system to the diarized speech segments by STHG produces a competitive performance on the Speech Transcription task of this challenge.
Authors: Fan Liu, Delong Chen, Zhangqingyun Guan, Xiaocong Zhou, Jiale Zhu, Jun Zhou
General-purpose foundation models have become increasingly important in the field of artificial intelligence. While self-supervised learning (SSL) and Masked Image Modeling (MIM) have led to promising results in building such foundation models for remote sensing, these models primarily learn low-level features, require annotated data for fine-tuning, and not applicable for retrieval and zero-shot applications due to the lack of language understanding. In response to these limitations, we propose RemoteCLIP, the first vision-language foundation model for remote sensing that aims to learn robust visual features with rich semantics, as well as aligned text embeddings for seamless downstream application. To address the scarcity of pre-training data, we leverage data scaling, converting heterogeneous annotations based on Box-to-Caption (B2C) and Mask-to-Box (M2B) conversion, and further incorporating UAV imagery, resulting a 12xlarger pretraining dataset. RemoteCLIP can be applied to a variety of downstream tasks, including zero-shot image classification, linear probing, k-NN classification, few-shot classification, image-text retrieval, and object counting. Evaluations on 16 datasets, including a newly introduced RemoteCount benchmark to test the object counting ability, show that RemoteCLIP consistently outperforms baseline foundation models across different model scales. Impressively, RemoteCLIP outperform previous SoTA by 9.14% mean recall on RSICD dataset and by 8.92% on RSICD dataset. For zero-shot classification, our RemoteCLIP outperform CLIP baseline by up to 6.39% average accuracy on 12 downstream datasets.Pretrained models is available at https://github.com/ChenDelong1999/RemoteCLIP .
Authors: Yuhang Huang, Zheng Qin, Xinwang Liu, Kai Xu
Recent diffusion probabilistic models (DPMs) have shown remarkable abilities of generated content, however, they often suffer from complex forward processes, resulting in inefficient solutions for the reversed process and prolonged sampling times. In this paper, we aim to address the aforementioned challenges by focusing on the diffusion process itself that we propose to decouple the intricate diffusion process into two comparatively simpler process to improve the generative efficacy and speed. In particular, we present a novel diffusion paradigm named DDM (Decoupled Diffusion Models) based on the Ito diffusion process, in which the image distribution is approximated by an explicit transition probability while the noise path is controlled by the standard Wiener process. We find that decoupling the diffusion process reduces the learning difficulty and the explicit transition probability improves the generative speed significantly. We prove a new training objective for DPM, which enables the model to learn to predict the noise and image components separately. Moreover, given the novel forward diffusion equation, we derive the reverse denoising formula of DDM that naturally supports fewer steps of generation without ordinary differential equation (ODE) based accelerators. Our experiments demonstrate that DDM outperforms previous DPMs by a large margin in fewer function evaluations setting and gets comparable performances in long function evaluations setting. We also show that our framework can be applied to image-conditioned generation and high-resolution image synthesis, and that it can generate high-quality images with only 10 function evaluations.
Authors: Shitao Tang, Fuyang Zhang, Jiacheng Chen, Peng Wang, Yasutaka Furukawa
This paper introduces MVDiffusion, a simple yet effective method for generating consistent multi-view images from text prompts given pixel-to-pixel correspondences (e.g., perspective crops from a panorama or multi-view images given depth maps and poses). Unlike prior methods that rely on iterative image warping and inpainting, MVDiffusion simultaneously generates all images with a global awareness, effectively addressing the prevalent error accumulation issue. At its core, MVDiffusion processes perspective images in parallel with a pre-trained text-to-image diffusion model, while integrating novel correspondence-aware attention layers to facilitate cross-view interactions. For panorama generation, while only trained with 10k panoramas, MVDiffusion is able to generate high-resolution photorealistic images for arbitrary texts or extrapolate one perspective image to a 360-degree view. For multi-view depth-to-image generation, MVDiffusion demonstrates state-of-the-art performance for texturing a scene mesh. The project page is at https://mvdiffusion.github.io/.
Authors: Zhuoxiao Chen, Yadan Luo, Zheng Wang, Mahsa Baktashmotlagh, Zi Huang
Unsupervised domain adaptation (DA) with the aid of pseudo labeling techniques has emerged as a crucial approach for domain-adaptive 3D object detection. While effective, existing DA methods suffer from a substantial drop in performance when applied to a multi-class training setting, due to the co-existence of low-quality pseudo labels and class imbalance issues. In this paper, we address this challenge by proposing a novel ReDB framework tailored for learning to detect all classes at once. Our approach produces Reliable, Diverse, and class-Balanced pseudo 3D boxes to iteratively guide the self-training on a distributionally different target domain. To alleviate disruptions caused by the environmental discrepancy (e.g., beam numbers), the proposed cross-domain examination (CDE) assesses the correctness of pseudo labels by copy-pasting target instances into a source environment and measuring the prediction consistency. To reduce computational overhead and mitigate the object shift (e.g., scales and point densities), we design an overlapped boxes counting (OBC) metric that allows to uniformly downsample pseudo-labeled objects across different geometric characteristics. To confront the issue of inter-class imbalance, we progressively augment the target point clouds with a class-balanced set of pseudo-labeled target instances and source objects, which boosts recognition accuracies on both frequently appearing and rare classes. Experimental results on three benchmark datasets using both voxel-based (i.e., SECOND) and point-based 3D detectors (i.e., PointRCNN) demonstrate that our proposed ReDB approach outperforms existing 3D domain adaptation methods by a large margin, improving 23.15% mAP on the nuScenes $\rightarrow$ KITTI task. The code is available at https://github.com/zhuoxiao-chen/ReDB-DA-3Ddet.
Authors: Dean Hazineh, Soon Wei Daniel Lim, Qi Guo, Federico Capasso, Todd Zickler
Optical metasurfaces composed of precisely engineered nanostructures have gained significant attention for their ability to manipulate light and implement distinct functionalities based on the properties of the incident field. Computational imaging systems have started harnessing this capability to produce sets of coded measurements that benefit certain tasks when paired with digital post-processing. Inspired by these works, we introduce a new system that uses a birefringent metasurface with a polarizer-mosaicked photosensor to capture four optically-coded measurements in a single exposure. We apply this system to the task of incoherent opto-electronic filtering, where digital spatial-filtering operations are replaced by simpler, per-pixel sums across the four polarization channels, independent of the spatial filter size. In contrast to previous work on incoherent opto-electronic filtering that can realize only one spatial filter, our approach can realize a continuous family of filters from a single capture, with filters being selected from the family by adjusting the post-capture digital summation weights. To find a metasurface that can realize a set of user-specified spatial filters, we introduce a form of gradient descent with a novel regularizer that encourages light efficiency and a high signal-to-noise ratio. We demonstrate several examples in simulation and with fabricated prototypes, including some with spatial filters that have prescribed variations with respect to depth and wavelength.
Visit the Project Page at https://deanhazineh.github.io/publications/Multi_Image_Synthesis/MIS_Home.html
Authors: Yiran Wang, Min Shi, Jiaqi Li, Zihao Huang, Zhiguo Cao, Jianming Zhang, Ke Xian, Guosheng Lin
Video depth estimation aims to infer temporally consistent depth. Some methods achieve temporal consistency by finetuning a single-image depth model during test time using geometry and re-projection constraints, which is inefficient and not robust. An alternative approach is to learn how to enforce temporal consistency from data, but this requires well-designed models and sufficient video depth data. To address these challenges, we propose a plug-and-play framework called Neural Video Depth Stabilizer (NVDS) that stabilizes inconsistent depth estimations and can be applied to different single-image depth models without extra effort. We also introduce a large-scale dataset, Video Depth in the Wild (VDW), which consists of 14,203 videos with over two million frames, making it the largest natural-scene video depth dataset to our knowledge. We evaluate our method on the VDW dataset as well as two public benchmarks and demonstrate significant improvements in consistency, accuracy, and efficiency compared to previous approaches. Our work serves as a solid baseline and provides a data foundation for learning-based video depth models. We will release our dataset and code for future research.
Authors: Jinheng Xie, Yuexiang Li, Yawen Huang, Haozhe Liu, Wentian Zhang, Yefeng Zheng, Mike Zheng Shou
Recent text-to-image diffusion models have demonstrated an astonishing capacity to generate high-quality images. However, researchers mainly studied the way of synthesizing images with only text prompts. While some works have explored using other modalities as conditions, considerable paired data, e.g., box/mask-image pairs, and fine-tuning time are required for nurturing models. As such paired data is time-consuming and labor-intensive to acquire and restricted to a closed set, this potentially becomes the bottleneck for applications in an open world. This paper focuses on the simplest form of user-provided conditions, e.g., box or scribble. To mitigate the aforementioned problem, we propose a training-free method to control objects and contexts in the synthesized images adhering to the given spatial conditions. Specifically, three spatial constraints, i.e., Inner-Box, Outer-Box, and Corner Constraints, are designed and seamlessly integrated into the denoising step of diffusion models, requiring no additional training and massive annotated layout data. Extensive results show that the proposed constraints can control what and where to present in the images while retaining the ability of the Stable Diffusion model to synthesize with high fidelity and diverse concept coverage. The code is publicly available at https://github.com/Sierkinhane/BoxDiff.
Authors: Yanjun Wang, Qingping Sun, Wenjia Wang, Jun Ling, Zhongang Cai, Rong Xie, Li Song
Human mesh reconstruction from a single image is challenging in the presence of occlusion, which can be caused by self, objects, or other humans. Existing methods either fail to separate human features accurately or lack proper supervision for feature completion. In this paper, we propose Dense Inpainting Human Mesh Recovery (DIMR), a two-stage method that leverages dense correspondence maps to handle occlusion. Our method utilizes a dense correspondence map to separate visible human features and completes human features on a structured UV map dense human with an attention-based feature completion module. We also design a feature inpainting training procedure that guides the network to learn from unoccluded features. We evaluate our method on several datasets and demonstrate its superior performance under heavily occluded scenarios compared to other methods. Extensive experiments show that our method obviously outperforms prior SOTA methods on heavily occluded images and achieves comparable results on the standard benchmarks (3DPW).
Authors: Thomas Manzini, Robin Murphy
This paper details the challenges in applying two computer vision systems, an EfficientDET supervised learning model and the unsupervised RX spectral classifier, to 98.9 GB of drone imagery from the Wu-Murad wilderness search and rescue (WSAR) effort in Japan and identifies 3 directions for future research. There have been at least 19 proposed approaches and 3 datasets aimed at locating missing persons in drone imagery, but only 3 approaches (2 unsupervised and 1 of an unknown structure) are referenced in the literature as having been used in an actual WSAR operation. Of these proposed approaches, the EfficientDET architecture and the unsupervised spectral RX classifier were selected as the most appropriate for this setting. The EfficientDET model was applied to the HERIDAL dataset and despite achieving performance that is statistically equivalent to the state-of-the-art, the model fails to translate to the real world in terms of false positives (e.g., identifying tree limbs and rocks as people), and false negatives (e.g., failing to identify members of the search team). The poor results in practice for algorithms that showed good results on datasets suggest 3 areas of future research: more realistic datasets for wilderness SAR, computer vision models that are capable of seamlessly handling the variety of imagery that can be collected during actual WSAR operations, and better alignment on performance measures.
Authors: Zun Wang, Jialu Li, Yicong Hong, Yi Wang, Qi Wu, Mohit Bansal, Stephen Gould, Hao Tan, Yu Qiao
Recent research in language-guided visual navigation has demonstrated a significant demand for the diversity of traversable environments and the quantity of supervision for training generalizable agents. To tackle the common data scarcity issue in existing vision-and-language navigation datasets, we propose an effective paradigm for generating large-scale data for learning, which applies 1200+ photo-realistic environments from HM3D and Gibson datasets and synthesizes 4.9 million instruction trajectory pairs using fully-accessible resources on the web. Importantly, we investigate the influence of each component in this paradigm on the agent's performance and study how to adequately apply the augmented data to pre-train and fine-tune an agent. Thanks to our large-scale dataset, the performance of an existing agent can be pushed up (+11% absolute with regard to previous SoTA) to a significantly new best of 80% single-run success rate on the R2R test split by simple imitation learning. The long-lasting generalization gap between navigating in seen and unseen environments is also reduced to less than 1% (versus 8% in the previous best method). Moreover, our paradigm also facilitates different models to achieve new state-of-the-art navigation results on CVDN, REVERIE, and R2R in continuous environments.
Authors: Qiufan Ji, Lin Wang, Cong Shi, Shengshan Hu, Yingying Chen, Lichao Sun
Deep Neural Networks (DNNs) for 3D point cloud recognition are vulnerable to adversarial examples, threatening their practical deployment. Despite the many research endeavors have been made to tackle this issue in recent years, the diversity of adversarial examples on 3D point clouds makes them more challenging to defend against than those on 2D images. For examples, attackers can generate adversarial examples by adding, shifting, or removing points. Consequently, existing defense strategies are hard to counter unseen point cloud adversarial examples. In this paper, we first establish a comprehensive, and rigorous point cloud adversarial robustness benchmark to evaluate adversarial robustness, which can provide a detailed understanding of the effects of the defense and attack methods. We then collect existing defense tricks in point cloud adversarial defenses and then perform extensive and systematic experiments to identify an effective combination of these tricks. Furthermore, we propose a hybrid training augmentation methods that consider various types of point cloud adversarial examples to adversarial training, significantly improving the adversarial robustness. By combining these tricks, we construct a more robust defense framework achieving an average accuracy of 83.45\% against various attacks, demonstrating its capability to enabling robust learners. Our codebase are open-sourced on: \url{https://github.com/qiufan319/benchmark_pc_attack.git}.
Authors: Anant Khandelwal
Large text-to-image diffusion models have achieved remarkable success in generating diverse, high-quality images. Additionally, these models have been successfully leveraged to edit input images by just changing the text prompt. But when these models are applied to videos, the main challenge is to ensure temporal consistency and coherence across frames. In this paper, we propose InFusion, a framework for zero-shot text-based video editing leveraging large pre-trained image diffusion models. Our framework specifically supports editing of multiple concepts with pixel-level control over diverse concepts mentioned in the editing prompt. Specifically, we inject the difference in features obtained with source and edit prompts from U-Net residual blocks of decoder layers. When these are combined with injected attention features, it becomes feasible to query the source contents and scale edited concepts along with the injection of unedited parts. The editing is further controlled in a fine-grained manner with mask extraction and attention fusion, which cut the edited part from the source and paste it into the denoising pipeline for the editing prompt. Our framework is a low-cost alternative to one-shot tuned models for editing since it does not require training. We demonstrated complex concept editing with a generalised image model (Stable Diffusion v1.5) using LoRA. Adaptation is compatible with all the existing image diffusion techniques. Extensive experimental results demonstrate the effectiveness of existing methods in rendering high-quality and temporally consistent videos.
Authors: Junzhou Chen, Qian Huang, Yulin Chen, Linyi Qian, Chengyuan Yu
Nucleus image segmentation is a crucial step in the analysis, pathological diagnosis, and classification, which heavily relies on the quality of nucleus segmentation. However, the complexity of issues such as variations in nucleus size, blurred nucleus contours, uneven staining, cell clustering, and overlapping cells poses significant challenges. Current methods for nucleus segmentation primarily rely on nuclear morphology or contour-based approaches. Nuclear morphology-based methods exhibit limited generalization ability and struggle to effectively predict irregular-shaped nuclei, while contour-based extraction methods face challenges in accurately segmenting overlapping nuclei. To address the aforementioned issues, we propose a dual-branch network using hybrid attention based residual U-blocks for nucleus instance segmentation. The network simultaneously predicts target information and target contours. Additionally, we introduce a post-processing method that combines the target information and target contours to distinguish overlapping nuclei and generate an instance segmentation image. Within the network, we propose a context fusion block (CF-block) that effectively extracts and merges contextual information from the network. Extensive quantitative evaluations are conducted to assess the performance of our method. Experimental results demonstrate the superior performance of the proposed method compared to state-of-the-art approaches on the BNS, MoNuSeg, CoNSeg, and CPM-17 datasets.
Authors: Zhongjie Duan, Lizhou You, Chengyu Wang, Cen Chen, Ziheng Wu, Weining Qian, Jun Huang
In recent years, diffusion models have emerged as the most powerful approach in image synthesis. However, applying these models directly to video synthesis presents challenges, as it often leads to noticeable flickering contents. Although recently proposed zero-shot methods can alleviate flicker to some extent, we still struggle to generate coherent videos. In this paper, we propose DiffSynth, a novel approach that aims to convert image synthesis pipelines to video synthesis pipelines. DiffSynth consists of two key components: a latent in-iteration deflickering framework and a video deflickering algorithm. The latent in-iteration deflickering framework applies video deflickering to the latent space of diffusion models, effectively preventing flicker accumulation in intermediate steps. Additionally, we propose a video deflickering algorithm, named patch blending algorithm, that remaps objects in different frames and blends them together to enhance video consistency. One of the notable advantages of DiffSynth is its general applicability to various video synthesis tasks, including text-guided video stylization, fashion video synthesis, image-guided video stylization, video restoring, and 3D rendering. In the task of text-guided video stylization, we make it possible to synthesize high-quality videos without cherry-picking. The experimental results demonstrate the effectiveness of DiffSynth. All videos can be viewed on our project page. Source codes will also be released.
Authors: A. Emin Orhan
This paper asks whether current self-supervised learning methods, if sufficiently scaled up, would be able to reach human-level visual object recognition capabilities with the same type and amount of visual experience humans learn from. Previous work on this question only considered the scaling of data size. Here, we consider the simultaneous scaling of data size, model size, and image resolution. We perform a scaling experiment with vision transformers up to 633M parameters in size (ViT-H/14) trained with up to 5K hours of human-like video data (long, continuous, mostly egocentric videos) with image resolutions of up to 476x476 pixels. The efficiency of masked autoencoders (MAEs) as a self-supervised learning algorithm makes it possible to run this scaling experiment on an unassuming academic budget. We find that it is feasible to reach human-level object recognition capacity at sub-human scales of model size, data size, and image size, if these factors are scaled up simultaneously. To give a concrete example, we estimate that a 2.5B parameter ViT model trained with 20K hours (2.3 years) of human-like video data with a spatial resolution of 952x952 pixels should be able to reach roughly human-level accuracy on ImageNet. Human-level competence is thus achievable for a fundamental perceptual capability from human-like perceptual experience (human-like in both amount and type) with extremely generic learning algorithms and architectures and without any substantive inductive biases.
Authors: Juncheng Li, Kaihang Pan, Zhiqi Ge, Minghe Gao, Hanwang Zhang, Wei Ji, Wenqiao Zhang, Tat-Seng Chua, Siliang Tang, Yueting Zhuang
Multimodal Large Language Models (MLLMs) have recently sparked significant interest, which demonstrates emergent capabilities to serve as a general-purpose model for various vision-language tasks. However, existing methods mainly focus on limited types of instructions with a single image as visual context, which hinders the widespread availability of MLLMs. In this paper, we introduce the I4 benchmark to comprehensively evaluate the instruction following ability on complicated interleaved vision-language instructions, which involve intricate image-text sequential context, covering a diverse range of scenarios (e.g., visually-rich webpages/textbooks, lecture slides, embodied dialogue). Systematic evaluation on our I4 benchmark reveals a common defect of existing methods: the Visual Prompt Generator (VPG) trained on image-captioning alignment objective tends to attend to common foreground information for captioning but struggles to extract specific information required by particular tasks. To address this issue, we propose a generic and lightweight controllable knowledge re-injection module, which utilizes the sophisticated reasoning ability of LLMs to control the VPG to conditionally extract instruction-specific visual information and re-inject it into the LLM. Further, we introduce an annotation-free cross-attention guided counterfactual image training strategy to methodically learn the proposed module by collaborating a cascade of foundation models. Enhanced by the proposed module and training strategy, we present Cheetor, a Transformer-based MLLM that can effectively handle a wide variety of interleaved vision-language instructions and achieves state-of-the-art zero-shot performance across all tasks of I4, without high-quality multimodal instruction tuning data. Cheetor also exhibits competitive performance compared with state-of-the-art instruction tuned models on MME benchmark.
Authors: Yifan Gao, Jinpeng Lin, Min Zhou, Chuanbin Liu, Hongtao Xie, Tiezheng Ge, Yuning Jiang
Text design is one of the most critical procedures in poster design, as it relies heavily on the creativity and expertise of humans to design text images considering the visual harmony and text-semantic. This study introduces TextPainter, a novel multimodal approach that leverages contextual visual information and corresponding text semantics to generate text images. Specifically, TextPainter takes the global-local background image as a hint of style and guides the text image generation with visual harmony. Furthermore, we leverage the language model and introduce a text comprehension module to achieve both sentence-level and word-level style variations. Besides, we construct the PosterT80K dataset, consisting of about 80K posters annotated with sentence-level bounding boxes and text contents. We hope this dataset will pave the way for further research on multimodal text image generation. Extensive quantitative and qualitative experiments demonstrate that TextPainter can generate visually-and-semantically-harmonious text images for posters.
Authors: Antonio Canela, Pol Caselles, Ibrar Malik, Eduard Ramon, Jaime García, Jordi Sánchez-Riera, Gil Triginer, Francesc Moreno-Noguer
Recent advances in full-head reconstruction have been obtained by optimizing a neural field through differentiable surface or volume rendering to represent a single scene. While these techniques achieve an unprecedented accuracy, they take several minutes, or even hours, due to the expensive optimization process required. In this work, we introduce InstantAvatar, a method that recovers full-head avatars from few images (down to just one) in a few seconds on commodity hardware. In order to speed up the reconstruction process, we propose a system that combines, for the first time, a voxel-grid neural field representation with a surface renderer. Notably, a naive combination of these two techniques leads to unstable optimizations that do not converge to valid solutions. In order to overcome this limitation, we present a novel statistical model that learns a prior distribution over 3D head signed distance functions using a voxel-grid based architecture. The use of this prior model, in combination with other design choices, results into a system that achieves 3D head reconstructions with comparable accuracy as the state-of-the-art with a 100x speed-up.
Authors: Tengchuan Kou, Xiaohong Liu, Wei Sun, Jun Jia, Xiongkuo Min, Guangtao Zhai, Ning Liu
Video shakiness is an unpleasant distortion of User Generated Content (UGC) videos, which is usually caused by the unstable hold of cameras. In recent years, many video stabilization algorithms have been proposed, yet no specific and accurate metric enables comprehensively evaluating the stability of videos. Indeed, most existing quality assessment models evaluate video quality as a whole without specifically taking the subjective experience of video stability into consideration. Therefore, these models cannot measure the video stability explicitly and precisely when severe shakes are present. In addition, there is no large-scale video database in public that includes various degrees of shaky videos with the corresponding subjective scores available, which hinders the development of Video Quality Assessment for Stability (VQA-S). To this end, we build a new database named StableDB that contains 1,952 diversely-shaky UGC videos, where each video has a Mean Opinion Score (MOS) on the degree of video stability rated by 34 subjects. Moreover, we elaborately design a novel VQA-S model named StableVQA, which consists of three feature extractors to acquire the optical flow, semantic, and blur features respectively, and a regression layer to predict the final stability score. Extensive experiments demonstrate that the StableVQA achieves a higher correlation with subjective opinions than the existing VQA-S models and generic VQA models. The database and codes are available at https://github.com/QMME/StableVQA.
Authors: Kai Huang, Feigege Wang, Ye Xi, Yutao Gao
Generalized Few-shot Semantic Segmentation (GFSS) extends Few-shot Semantic Segmentation (FSS) to simultaneously segment unseen classes and seen classes during evaluation. Previous works leverage additional branch or prototypical aggregation to eliminate the constrained setting of FSS. However, representation division and embedding prejudice, which heavily results in poor performance of GFSS, have not been synthetical considered. We address the aforementioned problems by jointing the prototypical kernel learning and open-set foreground perception. Specifically, a group of learnable kernels is proposed to perform segmentation with each kernel in charge of a stuff class. Then, we explore to merge the prototypical learning to the update of base-class kernels, which is consistent with the prototype knowledge aggregation of few-shot novel classes. In addition, a foreground contextual perception module cooperating with conditional bias based inference is adopted to perform class-agnostic as well as open-set foreground detection, thus to mitigate the embedding prejudice and prevent novel targets from being misclassified as background. Moreover, we also adjust our method to the Class Incremental Few-shot Semantic Segmentation (CIFSS) which takes the knowledge of novel classes in a incremental stream. Extensive experiments on PASCAL-5i and COCO-20i datasets demonstrate that our method performs better than previous state-of-the-art.
Authors: Fadi Boutros, Jonas Henry Grebe, Arjan Kuijper, Naser Damer
The availability of large-scale authentic face databases has been crucial to the significant advances made in face recognition research over the past decade. However, legal and ethical concerns led to the recent retraction of many of these databases by their creators, raising questions about the continuity of future face recognition research without one of its key resources. Synthetic datasets have emerged as a promising alternative to privacy-sensitive authentic data for face recognition development. However, recent synthetic datasets that are used to train face recognition models suffer either from limitations in intra-class diversity or cross-class (identity) discrimination, leading to less optimal accuracies, far away from the accuracies achieved by models trained on authentic data. This paper targets this issue by proposing IDiff-Face, a novel approach based on conditional latent diffusion models for synthetic identity generation with realistic identity variations for face recognition training. Through extensive evaluations, our proposed synthetic-based face recognition approach pushed the limits of state-of-the-art performances, achieving, for example, 98.00% accuracy on the Labeled Faces in the Wild (LFW) benchmark, far ahead from the recent synthetic-based face recognition solutions with 95.40% and bridging the gap to authentic-based face recognition with 99.82% accuracy.
Authors: Sneha Sree C, Mohammad Al Fahim, Keerthi Ram, Mohanasankar Sivaprakasam
Many segmentation networks have been proposed for 3D volumetric segmentation of tumors and organs at risk. Hospitals and clinical institutions seek to accelerate and minimize the efforts of specialists in image segmentation. Still, in case of errors generated by these networks, clinicians would have to manually edit the generated segmentation maps. Given a 3D volume and its putative segmentation map, we propose an approach to identify and measure erroneous regions in the segmentation map. Our method can estimate error at any point or node in a 3D mesh generated from a possibly erroneous volumetric segmentation map, serving as a Quality Assurance tool. We propose a graph neural network-based transformer based on the Nodeformer architecture to measure and classify the segmentation errors at any point. We have evaluated our network on a high-resolution micro-CT dataset of the human inner-ear bony labyrinth structure by simulating erroneous 3D segmentation maps. Our network incorporates a convolutional encoder to compute node-centric features from the input micro-CT data, the Nodeformer to learn the latent graph embeddings, and a Multi-Layer Perceptron (MLP) to compute and classify the node-wise errors. Our network achieves a mean absolute error of ~0.042 over other Graph Neural Networks (GNN) and an accuracy of 79.53% over other GNNs in estimating and classifying the node-wise errors, respectively. We also put forth vertex-normal prediction as a custom pretext task for pre-training the CNN encoder to improve the network's overall performance. Qualitative analysis shows the efficiency of our network in correctly classifying errors and reducing misclassifications.
Authors: Yifu Zhang, Hongru Li, Tao Yang, Rui Tao, Zhengyuan Liu, Shimeng Shi, Jiansong Zhang, Ning Ma, Wujin Feng, Zhanhu Zhang, Xinyu Zhang
Lesion segmentation of ultrasound medical images based on deep learning techniques is a widely used method for diagnosing diseases. Although there is a large amount of ultrasound image data in medical centers and other places, labeled ultrasound datasets are a scarce resource, and it is likely that no datasets are available for new tissues/organs. Transfer learning provides the possibility to solve this problem, but there are too many features in natural images that are not related to the target domain. As a source domain, redundant features that are not conducive to the task will be extracted. Migration between ultrasound images can avoid this problem, but there are few types of public datasets, and it is difficult to find sufficiently similar source domains. Compared with natural images, ultrasound images have less information, and there are fewer transferable features between different ultrasound images, which may cause negative transfer. To this end, a multi-source adversarial transfer learning network for ultrasound image segmentation is proposed. Specifically, to address the lack of annotations, the idea of adversarial transfer learning is used to adaptively extract common features between a certain pair of source and target domains, which provides the possibility to utilize unlabeled ultrasound data. To alleviate the lack of knowledge in a single source domain, multi-source transfer learning is adopted to fuse knowledge from multiple source domains. In order to ensure the effectiveness of the fusion and maximize the use of precious data, a multi-source domain independent strategy is also proposed to improve the estimation of the target domain data distribution, which further increases the learning ability of the multi-source adversarial migration learning network in multiple domains.