Authors: Tayssir Bouraffa, Elias Kjellberg Carlson, Erik Wessman, Ali Nouri, Pierre Lamart, Christian Berger
Abstract: Gathering data and identifying events in various traffic situations remains an essential challenge for the systematic evaluation of a perception system's performance. Analyzing large-scale, typically unstructured, multi-modal, time series data obtained from video, radar, and LiDAR is computationally demanding, particularly when meta-information or annotations are missing. We compare Optical Flow (OF) and Deep Learning (DL) to feed computationally efficient event detection via space-filling curves on video data from a forward-facing, in-vehicle camera. Our first approach leverages unexpected disturbances in the OF field from vehicle surroundings; the second approach is a DL model trained on human visual attention to predict a driver's gaze to spot potential event locations. We feed these results to a space-filling curve to reduce dimensionality and achieve computationally efficient event retrieval. We systematically evaluate our concept by obtaining characteristic patterns for both approaches from a large-scale virtual dataset (SMIRK) and applied our findings to the Zenseact Open Dataset (ZOD), a large multi-modal, real-world dataset, collected over two years in 14 different European countries. Our results yield that the OF approach excels in specificity and reduces false positives, while the DL approach demonstrates superior sensitivity. Both approaches offer comparable processing speed, making them suitable for real-time applications.
Authors: Chenxi Liu, Siqi Wang, Matthew Fisher, Deepali Aneja, Alec Jacobson
Abstract: Effective representation of 2D images is fundamental in digital image processing, where traditional methods like raster and vector graphics struggle with sharpness and textural complexity respectively. Current neural fields offer high-fidelity and resolution independence but require predefined meshes with known discontinuities, restricting their utility. We observe that by treating all mesh edges as potential discontinuities, we can represent the magnitude of discontinuities with continuous variables and optimize. Based on this observation, we introduce a novel discontinuous neural field model that jointly approximate the target image and recovers discontinuities. Through systematic evaluations, our neural field demonstrates superior performance in denoising and super-resolution tasks compared to InstantNGP, achieving improvements of over 5dB and 10dB, respectively. Our model also outperforms Mumford-Shah-based methods in accurately capturing discontinuities, with Chamfer distances 3.5x closer to the ground truth. Additionally, our approach shows remarkable capability in handling complex artistic drawings and natural images.
Authors: Weiheng Yao, Shuqiang Wang
Abstract: Multi-modal neuroimaging analysis is crucial for a comprehensive understanding of brain function and pathology, as it allows for the integration of different imaging techniques, thus overcoming the limitations of individual modalities. However, the high costs and limited availability of certain modalities pose significant challenges. To address these issues, this paper proposed the Condition-Aligned Temporal Diffusion (CATD) framework for end-to-end cross-modal synthesis of neuroimaging, enabling the generation of functional magnetic resonance imaging (fMRI)-detected Blood Oxygen Level Dependent (BOLD) signals from more accessible Electroencephalography (EEG) signals. By constructing Conditionally Aligned Block (CAB), heterogeneous neuroimages are aligned into a potential space, achieving a unified representation that provides the foundation for cross-modal transformation in neuroimaging. The combination with the constructed Dynamic Time-Frequency Segmentation (DTFS) module also enables the use of EEG signals to improve the temporal resolution of BOLD signals, thus augmenting the capture of the dynamic details of the brain. Experimental validation demonstrated the effectiveness of the framework in improving the accuracy of neural activity prediction, identifying abnormal brain regions, and enhancing the temporal resolution of BOLD signals. The proposed framework establishes a new paradigm for cross-modal synthesis of neuroimaging by unifying heterogeneous neuroimaging data into a potential representation space, showing promise in medical applications such as improving Parkinson's disease prediction and identifying abnormal brain regions.
Authors: Clemens Otte, Yinchong Yang, Danny Benlin Oswan
Abstract: The paper proposes a new testing approach for Deep Neural Networks (DNN) using gradient-free optimization to find perturbation chains that successfully falsify the tested DNN, going beyond existing grid-based or combinatorial testing. Applying it to an image segmentation task of detecting railway tracks in images, we demonstrate that the approach can successfully identify weaknesses of the tested DNN regarding particular combinations of common perturbations (e.g., rain, fog, blur, noise) on specific clusters of test images.
Authors: Sabah Abdulazeez Jebur, Khalid A. Hussein, Haider Kadhim Hoomod, Laith Alzubaidi, Ahmed Ali Saihood, YuanTong Gu
Abstract: Anomaly detection in videos is challenging due to the complexity, noise, and diverse nature of activities such as violence, shoplifting, and vandalism. While deep learning (DL) has shown excellent performance in this area, existing approaches have struggled to apply DL models across different anomaly tasks without extensive retraining. This repeated retraining is time-consuming, computationally intensive, and unfair. To address this limitation, a new DL framework is introduced in this study, consisting of three key components: transfer learning to enhance feature generalization, model fusion to improve feature representation, and multi-task classification to generalize the classifier across multiple tasks without training from scratch when new task is introduced. The framework's main advantage is its ability to generalize without requiring retraining from scratch for each new task. Empirical evaluations demonstrate the framework's effectiveness, achieving an accuracy of 97.99% on the RLVS dataset (violence detection), 83.59% on the UCF dataset (shoplifting detection), and 88.37% across both datasets using a single classifier without retraining. Additionally, when tested on an unseen dataset, the framework achieved an accuracy of 87.25%. The study also utilizes two explainability tools to identify potential biases, ensuring robustness and fairness. This research represents the first successful resolution of the generalization issue in anomaly detection, marking a significant advancement in the field.
Authors: Jiayuan Zhu, Yunli Qi, Junde Wu
Abstract: In this paper, we introduce Medical SAM 2 (MedSAM-2), an advanced segmentation model that utilizes the SAM 2 framework to address both 2D and 3D medical image segmentation tasks. By adopting the philosophy of taking medical images as videos, MedSAM-2 not only applies to 3D medical images but also unlocks new One-prompt Segmentation capability. That allows users to provide a prompt for just one or a specific image targeting an object, after which the model can autonomously segment the same type of object in all subsequent images, regardless of temporal relationships between the images. We evaluated MedSAM-2 across a variety of medical imaging modalities, including abdominal organs, optic discs, brain tumors, thyroid nodules, and skin lesions, comparing it against state-of-the-art models in both traditional and interactive segmentation settings. Our findings show that MedSAM-2 not only surpasses existing models in performance but also exhibits superior generalization across a range of medical image segmentation tasks. Our code will be released at: https://github.com/MedicineToken/Medical-SAM2
Authors: R\'ois\'in Luo, Alexandru Drimbarean, James McDermott, Colm O'Riordan
Abstract: This paper explores a novel paradigm in low-bit (i.e. 4-bits or lower) quantization, differing from existing state-of-the-art methods, by framing optimal quantization as an architecture search problem within convolutional neural networks (ConvNets). Our framework, dubbed \textbf{CoRa} (Optimal Quantization Residual \textbf{Co}nvolutional Operator Low-\textbf{Ra}nk Adaptation), is motivated by two key aspects. Firstly, quantization residual knowledge, i.e. the lost information between floating-point weights and quantized weights, has long been neglected by the research community. Reclaiming the critical residual knowledge, with an infinitesimal extra parameter cost, can reverse performance degradation without training. Secondly, state-of-the-art quantization frameworks search for optimal quantized weights to address the performance degradation. Yet, the vast search spaces in weight optimization pose a challenge for the efficient optimization in large models. For example, state-of-the-art BRECQ necessitates $2 \times 10^4$ iterations to quantize models. Fundamentally differing from existing methods, \textbf{CoRa} searches for the optimal architectures of low-rank adapters, reclaiming critical quantization residual knowledge, within the search spaces smaller compared to the weight spaces, by many orders of magnitude. The low-rank adapters approximate the quantization residual weights, discarded in previous methods. We evaluate our approach over multiple pre-trained ConvNets on ImageNet. \textbf{CoRa} achieves comparable performance against both state-of-the-art quantization-aware training and post-training quantization baselines, in $4$-bit and $3$-bit quantization, by using less than $250$ iterations on a small calibration set with $1600$ images. Thus, \textbf{CoRa} establishes a new state-of-the-art in terms of the optimization efficiency in low-bit quantization.
Authors: Bin Han, Yiwei Yang, Anat Caspi, Bill Howe
Abstract: Equitable urban transportation applications require high-fidelity digital representations of the built environment: not just streets and sidewalks, but bike lanes, marked and unmarked crossings, curb ramps and cuts, obstructions, traffic signals, signage, street markings, potholes, and more. Direct inspections and manual annotations are prohibitively expensive at scale. Conventional machine learning methods require substantial annotated training data for adequate performance. In this paper, we consider vision language models as a mechanism for annotating diverse urban features from satellite images, reducing the dependence on human annotation to produce large training sets. While these models have achieved impressive results in describing common objects in images captured from a human perspective, their training sets are less likely to include strong signals for esoteric features in the built environment, and their performance in these settings is therefore unclear. We demonstrate proof-of-concept combining a state-of-the-art vision language model and variants of a prompting strategy that asks the model to consider segmented elements independently of the original image. Experiments on two urban features -- stop lines and raised tables -- show that while direct zero-shot prompting correctly annotates nearly zero images, the pre-segmentation strategies can annotate images with near 40% intersection-over-union accuracy. We describe how these results inform a new research agenda in automatic annotation of the built environment to improve equity, accessibility, and safety at broad scale and in diverse environments.
Authors: Chengbo Zang, Mehmet Kerem Turkcan, Gil Zussman, Javad Ghaderi, Zoran Kostic
Abstract: We present a novel data-driven simulation environment for modeling traffic in metropolitan street intersections. Using real-world tracking data collected over an extended period of time, we train trajectory forecasting models to learn agent interactions and environmental constraints that are difficult to capture conventionally. Trajectories of new agents are first coarsely generated by sampling from the spatial and temporal generative distributions, then refined using state-of-the-art trajectory forecasting models. The simulation can run either autonomously, or under explicit human control conditioned on the generative distributions. We present the experiments for a variety of model configurations. Under an iterative prediction scheme, the way-point-supervised TrajNet++ model obtained 0.36 Final Displacement Error (FDE) in 20 FPS on an NVIDIA A100 GPU.
Authors: Lingyu Du, Jinyuan Jia, Xucong Zhang, Guohao Lan
Abstract: Eye gaze contains rich information about human attention and cognitive processes. This capability makes the underlying technology, known as gaze tracking, a critical enabler for many ubiquitous applications and has triggered the development of easy-to-use gaze estimation services. Indeed, by utilizing the ubiquitous cameras on tablets and smartphones, users can readily access many gaze estimation services. In using these services, users must provide their full-face images to the gaze estimator, which is often a black box. This poses significant privacy threats to the users, especially when a malicious service provider gathers a large collection of face images to classify sensitive user attributes. In this work, we present PrivateGaze, the first approach that can effectively preserve users' privacy in black-box gaze tracking services without compromising gaze estimation performance. Specifically, we proposed a novel framework to train a privacy preserver that converts full-face images into obfuscated counterparts, which are effective for gaze estimation while containing no privacy information. Evaluation on four datasets shows that the obfuscated image can protect users' private information, such as identity and gender, against unauthorized attribute classification. Meanwhile, when used directly by the black-box gaze estimator as inputs, the obfuscated images lead to comparable tracking performance to the conventional, unprotected full-face images.
Authors: Mohammed Rakib, Adil Aman Mohammed, Cole Diggins, Sumit Sharma, Jeff Michael Sadler, Tyson Ochsner, Arun Bagavathi
Abstract: Soil moisture estimation is an important task to enable precision agriculture in creating optimal plans for irrigation, fertilization, and harvest. It is common to utilize statistical and machine learning models to estimate soil moisture from traditional data sources such as weather forecasts, soil properties, and crop properties. However, there is a growing interest in utilizing aerial and geospatial imagery to estimate soil moisture. Although these images capture high-resolution crop details, they are expensive to curate and challenging to interpret. Imagine, an AI-enhanced software tool that predicts soil moisture using visual cues captured by smartphones and statistical data given by weather forecasts. This work is a first step towards that goal of developing a multi-modal approach for soil moisture estimation. In particular, we curate a dataset consisting of real-world images taken from ground stations and their corresponding weather data. We also propose MIS-ME - Meteorological & Image based Soil Moisture Estimator, a multi-modal framework for soil moisture estimation. Our extensive analysis shows that MIS-ME achieves a MAPE of 10.79%, outperforming traditional unimodal approaches with a reduction of 2.6% in MAPE for meteorological data and 1.5% in MAPE for image data, highlighting the effectiveness of tailored multi-modal approaches.
Authors: Jesus Guerrero
Abstract: This work shows a procedural method for extracting object heights from LiDAR and aerial imagery. We discuss how to get heights and the future of LiDAR and imagery processing. SOTA object segmentation allows us to take get object heights with no deep learning background. Engineers will be keeping track of world data across generations and reprocessing them. They will be using older procedural methods like this paper and newer ones discussed here. SOTA methods are going beyond analysis and into generative AI. We cover both a procedural methodology and the newer ones performed with language models. These include point cloud, imagery and text encoding allowing for spatially aware AI.
Authors: Yabin Zhu, Qianwu Wang, Chenglong Li, Jin Tang, Zhixiang Huang
Abstract: The complementary benefits from visible and thermal infrared data are widely utilized in various computer vision task, such as visual tracking, semantic segmentation and object detection, but rarely explored in Multiple Object Tracking (MOT). In this work, we contribute a large-scale Visible-Thermal video benchmark for MOT, called VT-MOT. VT-MOT has the following main advantages. 1) The data is large scale and high diversity. VT-MOT includes 582 video sequence pairs, 401k frame pairs from surveillance, drone, and handheld platforms. 2) The cross-modal alignment is highly accurate. We invite several professionals to perform both spatial and temporal alignment frame by frame. 3) The annotation is dense and high-quality. VT-MOT has 3.99 million annotation boxes annotated and double-checked by professionals, including heavy occlusion and object re-acquisition (object disappear and reappear) challenges. To provide a strong baseline, we design a simple yet effective tracking framework, which effectively fuses temporal information and complementary information of two modalities in a progressive manner, for robust visible-thermal MOT. A comprehensive experiment are conducted on VT-MOT and the results prove the superiority and effectiveness of the proposed method compared with state-of-the-art methods. From the evaluation results and analysis, we specify several potential future directions for visible-thermal MOT. The project is released in https://github.com/wqw123wqw/PFTrack.
Authors: Xiang Gao, Jiaying Liu
Abstract: Large-scale text-to-image diffusion models have been a revolutionary milestone in the evolution of generative AI and multimodal technology, allowing extraordinary image generation based on natural-language text prompts. However, the issue of lacking controllability of such models restricts their practical applicability for real-life content creation, for which attention has been focused on leveraging a reference image to control text-to-image synthesis. Due to the close correlation between the reference image and the generated image, this problem can also be regarded as the task of manipulating (or editing) the reference image as per the text, namely text-driven image-to-image translation. This paper contributes a novel, concise, and efficient approach that adapts the pre-trained large-scale text-to-image (T2I) diffusion model to the image-to-image (I2I) paradigm in a plug-and-play manner, realizing high-quality and versatile text-driven I2I translation without any model training, model fine-tuning, or online optimization process. To guide T2I generation with a reference image, we propose to model diverse guiding factors with correspondingly different frequency bands of diffusion features in the DCT spectral space, and accordingly devise a novel frequency band substitution layer that dynamically substitutes a certain DCT frequency band of the diffusion features with the corresponding counterpart of the reference image along the reverse sampling process. We demonstrate that our method flexibly enables highly controllable text-driven I2I translation both in the guiding factor and guiding intensity of the reference image, simply by tuning the type and bandwidth of the substituted frequency band, respectively. Extensive qualitative and quantitative experiments verify the superiority of our approach over related methods in I2I translation visual quality, versatility, and controllability.
Authors: Die Chen, Zhiwen Li, Mingyuan Fan, Cen Chen, Wenmeng Zhou, Yaliang Li
Abstract: Text-to-image diffusion models have shown the ability to learn a diverse range of concepts. However, it is worth noting that they may also generate undesirable outputs, consequently giving rise to significant security concerns. Specifically, issues such as Not Safe for Work (NSFW) content and potential violations of style copyright may be encountered. Since image generation is conditioned on text, prompt purification serves as a straightforward solution for content safety. Similar to the approach taken by LLM, some efforts have been made to control the generation of safe outputs by purifying prompts. However, it is also important to note that even with these efforts, non-toxic text still carries a risk of generating non-compliant images, which is referred to as implicit unsafe prompts. Furthermore, some existing works fine-tune the models to erase undesired concepts from model weights. This type of method necessitates multiple training iterations whenever the concept is updated, which can be time-consuming and may potentially lead to catastrophic forgetting. To address these challenges, we propose a simple yet effective approach that incorporates non-compliant concepts into an erasure prompt. This erasure prompt proactively participates in the fusion of image spatial features and text embeddings. Through attention mechanisms, our method is capable of identifying feature representations of non-compliant concepts in the image space. We re-weight these features to effectively suppress the generation of unsafe images conditioned on original implicit unsafe prompts. Our method exhibits superior erasure effectiveness while achieving high scores in image fidelity compared to the state-of-the-art baselines. WARNING: This paper contains model outputs that may be offensive.
Authors: Yingying Zhang, Xin Guo, Jiangwei Lao, Lei Yu, Lixiang Ru, Jian Wang, Guo Ye, Huimei He, Jingdong Chen, Ming Yang
Abstract: Large-scale self-supervised pre-training has paved the way for one foundation model to handle many different vision tasks. Most pre-training methodologies train a single model of a certain size at one time. Nevertheless, various computation or storage constraints in real-world scenarios require substantial efforts to develop a series of models with different sizes to deploy. Thus, in this study, we propose a novel tri-branch self-supervised training framework, termed as POA (Pre-training Once for All), to tackle this aforementioned issue. Our approach introduces an innovative elastic student branch into a modern self-distillation paradigm. At each pre-training step, we randomly sample a sub-network from the original student to form the elastic student and train all branches in a self-distilling fashion. Once pre-trained, POA allows the extraction of pre-trained models of diverse sizes for downstream tasks. Remarkably, the elastic student facilitates the simultaneous pre-training of multiple models with different sizes, which also acts as an additional ensemble of models of various sizes to enhance representation learning. Extensive experiments, including k-nearest neighbors, linear probing evaluation and assessments on multiple downstream tasks demonstrate the effectiveness and advantages of our POA. It achieves state-of-the-art performance using ViT, Swin Transformer and ResNet backbones, producing around a hundred models with different sizes through a single pre-training session. The code is available at: https://github.com/Qichuzyy/POA.
Authors: Xiangbo Gao, Asiegbu Miracle Kanu-Asiegbu, Xiaoxiao Du
Abstract: This paper proposes MambaST, a plug-and-play cross-spectral spatial-temporal fusion pipeline for efficient pedestrian detection. Several challenges exist for pedestrian detection in autonomous driving applications. First, it is difficult to perform accurate detection using RGB cameras under dark or low-light conditions. Cross-spectral systems must be developed to integrate complementary information from multiple sensor modalities, such as thermal and visible cameras, to improve the robustness of the detections. Second, pedestrian detection models are latency-sensitive. Efficient and easy-to-scale detection models with fewer parameters are highly desirable for real-time applications such as autonomous driving. Third, pedestrian video data provides spatial-temporal correlations of pedestrian movement. It is beneficial to incorporate temporal as well as spatial information to enhance pedestrian detection. This work leverages recent advances in the state space model (Mamba) and proposes a novel Multi-head Hierarchical Patching and Aggregation (MHHPA) structure to extract both fine-grained and coarse-grained information from both RGB and thermal imagery. Experimental results show that the proposed MHHPA is an effective and efficient alternative to a Transformer model for cross-spectral pedestrian detection. Our proposed model also achieves superior performance on small-scale pedestrian detection. The code is available at https://github.com/XiangboGaoBarry/MambaST}{https://github.com/XiangboGaoBarry/MambaST.
URLs: https://github.com/XiangboGaoBarry/MambaST, https://github.com/XiangboGaoBarry/MambaST.
Authors: Yang Jin, Lei Zhang, Shi Yan, Bin Fan, Binglu Wang
Abstract: Gaze object prediction (GOP) aims to predict the category and location of the object that a human is looking at. Previous methods utilized box-level supervision to identify the object that a person is looking at, but struggled with semantic ambiguity, ie, a single box may contain several items since objects are close together. The Vision foundation model (VFM) has improved in object segmentation using box prompts, which can reduce confusion by more precisely locating objects, offering advantages for fine-grained prediction of gaze objects. This paper presents a more challenging gaze object segmentation (GOS) task, which involves inferring the pixel-level mask corresponding to the object captured by human gaze behavior. In particular, we propose that the pixel-level supervision provided by VFM can be integrated into gaze object prediction to mitigate semantic ambiguity. This leads to our gaze object detection and segmentation framework that enables accurate pixel-level predictions. Different from previous methods that require additional head input or ignore head features, we propose to automatically obtain head features from scene features to ensure the model's inference efficiency and flexibility in the real world. Moreover, rather than directly fuse features to predict gaze heatmap as in existing methods, which may overlook spatial location and subtle details of the object, we develop a space-to-object gaze regression method to facilitate human-object gaze interaction. Specifically, it first constructs an initial human-object spatial connection, then refines this connection by interacting with semantically clear features in the segmentation branch, ultimately predicting a gaze heatmap for precise localization. Extensive experiments on GOO-Synth and GOO-Real datasets demonstrate the effectiveness of our method.
Authors: Ruohua Shi, Zhaochen Liu, Lingyu Duan, Tingting Jiang
Abstract: Segmentation of surgical instruments is crucial for enhancing surgeon performance and ensuring patient safety. Conventional techniques such as binary, semantic, and instance segmentation share a common drawback: they do not accommodate the parts of instruments obscured by tissues or other instruments. Precisely predicting the full extent of these occluded instruments can significantly improve laparoscopic surgeries by providing critical guidance during operations and assisting in the analysis of potential surgical errors, as well as serving educational purposes. In this paper, we introduce Amodal Segmentation to the realm of surgical instruments in the medical field. This technique identifies both the visible and occluded parts of an object. To achieve this, we introduce a new Amoal Instruments Segmentation (AIS) dataset, which was developed by reannotating each instrument with its complete mask, utilizing the 2017 MICCAI EndoVis Robotic Instrument Segmentation Challenge dataset. Additionally, we evaluate several leading amodal segmentation methods to establish a benchmark for this new dataset.
Authors: Lu Yu, Zhe Tao, Hantao Yao, Joost Van de Weijer, Changsheng Xu
Abstract: Deep neural networks (DNNs) excel on fixed datasets but struggle with incremental and shifting data in real-world scenarios. Continual learning addresses this challenge by allowing models to learn from new data while retaining previously learned knowledge. Existing methods mainly rely on visual features, often neglecting the rich semantic information encoded in text. The semantic knowledge available in the label information of the images, offers important semantic information that can be related with previously acquired knowledge of semantic classes. Consequently, effectively leveraging this information throughout continual learning is expected to be beneficial. To address this, we propose integrating semantic guidance within and across tasks by capturing semantic similarity using text embeddings. We start from a pre-trained CLIP model, employ the \emph{Semantically-guided Representation Learning (SG-RL)} module for a soft-assignment towards all current task classes, and use the Semantically-guided Knowledge Distillation (SG-KD) module for enhanced knowledge transfer. Experimental results demonstrate the superiority of our method on general and fine-grained datasets. Our code can be found in https://github.com/aprilsveryown/semantically-guided-continual-learning.
URLs: https://github.com/aprilsveryown/semantically-guided-continual-learning.
Authors: Zhixin Yan, Yan Zhong, Wenjun Zhang, Lin Shu, Hongbin Xu, Wenxiong Kang
Abstract: Remote Photoplethysmography (rPPG) is a non-contact technique for extracting physiological signals from facial videos, used in applications like emotion monitoring, medical assistance, and anti-face spoofing. Unlike controlled laboratory settings, real-world environments often contain motion artifacts and noise, affecting the performance of existing methods. To address this, we propose PhysMamba, a dual-stream time-frequency interactive model based on Mamba. PhysMamba integrates the state-of-the-art Mamba-2 model and employs a dual-stream architecture to learn diverse rPPG features, enhancing robustness in noisy conditions. Additionally, we designed the Cross-Attention State Space Duality (CASSD) module to improve information exchange and feature complementarity between the two streams. We validated PhysMamba using PURE, UBFC-rPPG and MMPD. Experimental results show that PhysMamba achieves state-of-the-art performance across various scenarios, particularly in complex environments, demonstrating its potential in practical remote heart rate monitoring applications.
Authors: Hesong Li, Ying Fu
Abstract: Visible and infrared image fusion (VIF) aims to combine information from visible and infrared images into a single fused image. Previous VIF methods usually employ a color space transformation to keep the hue and saturation from the original visible image. However, for fast VIF methods, this operation accounts for the majority of the calculation and is the bottleneck preventing faster processing. In this paper, we propose a fast fusion method, FCDFusion, with little color deviation. It preserves color information without color space transformations, by directly operating in RGB color space. It incorporates gamma correction at little extra cost, allowing color and contrast to be rapidly improved. We regard the fusion process as a scaling operation on 3D color vectors, greatly simplifying the calculations. A theoretical analysis and experiments show that our method can achieve satisfactory results in only 7 FLOPs per pixel. Compared to state-of-the-art fast, color-preserving methods using HSV color space, our method provides higher contrast at only half of the computational cost. We further propose a new metric, color deviation, to measure the ability of a VIF method to preserve color. It is specifically designed for VIF tasks with color visible-light images, and overcomes deficiencies of existing VIF metrics used for this purpose. Our code is available at https://github.com/HeasonLee/FCDFusion.
Authors: Ajinkya Shinde, Gaurav Sharma, Manisha Pattanaik, Sri Niwas Singh
Abstract: LiDAR-based sensors employing optical spectrum signals play a vital role in providing significant information about the target objects in autonomous driving vehicle systems. However, the presence of fog in the atmosphere severely degrades the overall system's performance. This manuscript analyzes the role of fog particle size distributions in 3D object detection under adverse weather conditions. We utilise Mie theory and meteorological optical range (MOR) to calculate the attenuation and backscattering coefficient values for point cloud generation and analyze the overall system's accuracy in Car, Cyclist, and Pedestrian case scenarios under easy, medium and hard detection difficulties. Gamma and Junge (Power-Law) distributions are employed to mathematically model the fog particle size distribution under strong and moderate advection fog environments. Subsequently, we modified the KITTI dataset based on the backscattering coefficient values and trained it on the PV-RCNN++ deep neural network model for Car, Cyclist, and Pedestrian cases under different detection difficulties. The result analysis shows a significant variation in the system's accuracy concerning the changes in target object dimensionality, the nature of the fog environment and increasing detection difficulties, with the Car exhibiting the highest accuracy of around 99% and the Pedestrian showing the lowest accuracy of around 73%.
Authors: Yucheng Yang, Xiang Gu, Jian Sun
Abstract: Universal domain adaptation (UniDA) aims to transfer knowledge from a labeled source domain to an unlabeled target domain without requiring the same label sets of both domains. The existence of domain and category shift makes the task challenging and requires us to distinguish "known" samples (i.e., samples whose labels exist in both domains) and "unknown" samples (i.e., samples whose labels exist in only one domain) in both domains before reducing the domain gap. In this paper, we consider the problem from the point of view of distribution matching which we only need to align two distributions partially. A novel approach, dubbed mini-batch Prototypical Partial Optimal Transport (m-PPOT), is proposed to conduct partial distribution alignment for UniDA. In training phase, besides minimizing m-PPOT, we also leverage the transport plan of m-PPOT to reweight source prototypes and target samples, and design reweighted entropy loss and reweighted cross-entropy loss to distinguish "known" and "unknown" samples. Experiments on four benchmarks show that our method outperforms the previous state-of-the-art UniDA methods.
Authors: Donwon Park, Hayeon Kim, Se Young Chun
Abstract: Recently, pre-trained model and efficient parameter tuning have achieved remarkable success in natural language processing and high-level computer vision with the aid of masked modeling and prompt tuning. In low-level computer vision, however, there have been limited investigations on pre-trained models and even efficient fine-tuning strategy has not yet been explored despite its importance and benefit in various real-world tasks such as alleviating memory inflation issue when integrating new tasks on AI edge devices. Here, we propose a novel efficient parameter tuning approach dubbed contribution-based low-rank adaptation (CoLoRA) for multiple image restorations along with effective pre-training method with random order degradations (PROD). Unlike prior arts that tune all network parameters, our CoLoRA effectively fine-tunes small amount of parameters by leveraging LoRA (low-rank adaptation) for each new vision task with our contribution-based method to adaptively determine layer by layer capacity for that task to yield comparable performance to full tuning. Furthermore, our PROD strategy allows to extend the capability of pre-trained models with improved performance as well as robustness to bridge synthetic pre-training and real-world fine-tuning. Our CoLoRA with PROD has demonstrated its superior performance in various image restoration tasks across diverse degradation types on both synthetic and real-world datasets for known and novel tasks.
Authors: Wei Chen, Long Chen, Yu Wu
Abstract: Most advanced visual grounding methods rely on Transformers for visual-linguistic feature fusion. However, these Transformer-based approaches encounter a significant drawback: the computational costs escalate quadratically due to the self-attention mechanism in the Transformer Encoder, particularly when dealing with high-resolution images or long context sentences. This quadratic increase in computational burden restricts the applicability of visual grounding to more intricate scenes, such as conversation-based reasoning segmentation, which involves lengthy language expressions. In this paper, we propose an efficient and effective multi-task visual grounding (EEVG) framework based on Transformer Decoder to address this issue, which reduces the cost in both language and visual aspects. In the language aspect, we employ the Transformer Decoder to fuse visual and linguistic features, where linguistic features are input as memory and visual features as queries. This allows fusion to scale linearly with language expression length. In the visual aspect, we introduce a parameter-free approach to reduce computation by eliminating background visual tokens based on attention scores. We then design a light mask head to directly predict segmentation masks from the remaining sparse feature maps. Extensive results and ablation studies on benchmarks demonstrate the efficiency and effectiveness of our approach. Code is available in https://github.com/chenwei746/EEVG.
Authors: Furkan Aykut Sarikamis, Abdullah Aydin Alatan
Abstract: 3D Gaussian Splatting has recently shown promising results as an alternative scene representation in SLAM systems to neural implicit representations. However, current methods either lack dense depth maps to supervise the mapping process or detailed training designs that consider the scale of the environment. To address these drawbacks, we present IG-SLAM, a dense RGB-only SLAM system that employs robust Dense-SLAM methods for tracking and combines them with Gaussian Splatting. A 3D map of the environment is constructed using accurate pose and dense depth provided by tracking. Additionally, we utilize depth uncertainty in map optimization to improve 3D reconstruction. Our decay strategy in map optimization enhances convergence and allows the system to run at 10 fps in a single process. We demonstrate competitive performance with state-of-the-art RGB-only SLAM systems while achieving faster operation speeds. We present our experiments on the Replica, TUM-RGBD, ScanNet, and EuRoC datasets. The system achieves photo-realistic 3D reconstruction in large-scale sequences, particularly in the EuRoC dataset.
Authors: Changqun Xia, Chenxi Xie, Zhentao He, Tianshu Yu, Jia Li
Abstract: We present an advanced study on more challenging high-resolution salient object detection (HRSOD) from both dataset and network framework perspectives. To compensate for the lack of HRSOD dataset, we thoughtfully collect a large-scale high resolution salient object detection dataset, called UHRSD, containing 5,920 images from real-world complex scenarios at 4K-8K resolutions. All the images are finely annotated in pixel-level, far exceeding previous low-resolution SOD datasets. Aiming at overcoming the contradiction between the sampling depth and the receptive field size in the past methods, we propose a novel one-stage framework for HR-SOD task using pyramid grafting mechanism. In general, transformer-based and CNN-based backbones are adopted to extract features from different resolution images independently and then these features are grafted from transformer branch to CNN branch. An attention-based Cross-Model Grafting Module (CMGM) is proposed to enable CNN branch to combine broken detailed information more holistically, guided by different source feature during decoding process. Moreover, we design an Attention Guided Loss (AGL) to explicitly supervise the attention matrix generated by CMGM to help the network better interact with the attention from different branches. Comprehensive experiments on UHRSD and widely-used SOD datasets demonstrate that our method can simultaneously locate salient object and preserve rich details, outperforming state-of-the-art methods. To verify the generalization ability of the proposed framework, we apply it to the camouflaged object detection (COD) task. Notably, our method performs superior to most state-of-the-art COD methods without bells and whistles.
Authors: Farukh Yaushev, Daria Nogina, Valentin Samokhin, Mariya Dugova, Ekaterina Petrash, Dmitry Sevryukov, Mikhail Belyaev, Maxim Pisov
Abstract: Understanding body part geometry is crucial for precise medical diagnostics. Curves effectively describe anatomical structures and are widely used in medical imaging applications related to cardiovascular, respiratory, and skeletal diseases. Traditional curve detection methods are often task-specific, relying heavily on domain-specific features, limiting their broader applicability. This paper introduces a novel approach for detecting non-branching curves, which does not require prior knowledge of the object's orientation, shape, or position. Our method uses neural networks to predict (1) an attraction field, which offers subpixel accuracy, and (2) a closeness map, which limits the region of interest and essentially eliminates outliers far from the desired curve. We tested our curve detector on several clinically relevant tasks with diverse morphologies and achieved impressive subpixel-level accuracy results that surpass existing methods, highlighting its versatility and robustness. Additionally, to support further advancements in this field, we provide our private annotations of aortic centerlines and masks, which can serve as a benchmark for future research. The dataset can be found at https://github.com/neuro-ml/curve-detection.
Authors: Bryan Wong, Mun Yong Yi
Abstract: The classification of gigapixel-sized whole slide images (WSIs), digital representations of histological slides obtained via a high-resolution scanner, faces significant challenges associated with the meticulous and time-consuming nature of fine-grained labeling. While weakly-supervised multiple instance learning (MIL) has emerged as a promising approach, current MIL methods are constrained by their limited ability to leverage the wealth of information embedded within unlabeled WSIs. This limitation often necessitates training MIL feature aggregators from scratch after the feature extraction process, hindering efficiency and accuracy. PreMix extends the general MIL framework by pre-training the MIL aggregator with an intra-batch slide mixing approach. Specifically, PreMix incorporates Barlow Twins Slide Mixing during pre-training, enhancing its ability to handle diverse WSI sizes and maximizing the utility of unlabeled WSIs. Combined with Mixup and Manifold Mixup during fine-tuning, PreMix achieves a mean of 4.7% performance improvement over the baseline MIL framework, the hierarchical image pyramid transformer (HIPT), on the Camelyon16 dataset. The observed improvement across a range of active learning acquisition functions and WSI-labeled training budgets highlights the framework's adaptability to diverse datasets and varying resource constraints. Ultimately, PreMix paves the way for more efficient and accurate WSI classification under limited WSI-labeled datasets, encouraging the broader adoption of unlabeled WSI data in histopathological research. The code is available at https://anonymous.4open.science/r/PreMix
Authors: Bryan Wong, Mun Yong Yi
Abstract: Multiple instance learning (MIL) has become a preferred method for classifying gigapixel whole slide images (WSIs), without requiring patch label annotation. The focus of the current MIL research stream is on the embedding-based MIL approach, which involves extracting feature vectors from patches using a pre-trained feature extractor. These feature vectors are then fed into an MIL aggregator for slide-level prediction. Despite prior research suggestions on enhancing the most commonly used ResNet50 supervised model pre-trained on ImageNet-1K, there remains a lack of clear guidance on selecting the optimal feature extractor to maximize WSI performance. This study aims at addressing this gap by examining MIL feature extractors across three dimensions: pre-training dataset, backbone model, and pre-training method. Extensive experiments were carried out on the two public WSI datasets (TCGA-NSCLC and Camelyon16) using four SOTA MIL models. The main findings indicate the following: 1) Performance significantly improves with larger and more varied pre-training datasets in both CNN and Transformer backbones. 2) `Modern and deeper' backbones greatly outperform `standard' backbones (ResNet and ViT), with performance improvements more guaranteed in Transformer-based backbones. 3) The choice of self-supervised learning (SSL) method is crucial, with the most significant benefits observed when applied to the Transformer (ViT) backbone. The study findings have practical implications, including designing more effective pathological foundation models. Our code is available at: https://anonymous.4open.science/r/MIL-Feature-Extractor-Selection
URLs: https://anonymous.4open.science/r/MIL-Feature-Extractor-Selection
Authors: Qian Zhang, Xiangzi Dai, Ninghua Yang, Xiang An, Ziyong Feng, Xingyu Ren
Abstract: VAR is a new generation paradigm that employs 'next-scale prediction' as opposed to 'next-token prediction'. This innovative transformation enables auto-regressive (AR) transformers to rapidly learn visual distributions and achieve robust generalization. However, the original VAR model is constrained to class-conditioned synthesis, relying solely on textual captions for guidance. In this paper, we introduce VAR-CLIP, a novel text-to-image model that integrates Visual Auto-Regressive techniques with the capabilities of CLIP. The VAR-CLIP framework encodes captions into text embeddings, which are then utilized as textual conditions for image generation. To facilitate training on extensive datasets, such as ImageNet, we have constructed a substantial image-text dataset leveraging BLIP2. Furthermore, we delve into the significance of word positioning within CLIP for the purpose of caption guidance. Extensive experiments confirm VAR-CLIP's proficiency in generating fantasy images with high fidelity, textual congruence, and aesthetic excellence. Our project page are https://github.com/daixiangzi/VAR-CLIP
Authors: Ruitao Xie, Limai Jiang, Xiaoxi He, Yi Pan, Yunpeng Cai
Abstract: Machine-based brain tumor segmentation can help doctors make better diagnoses. However, the complex structure of brain tumors and expensive pixel-level annotations present challenges for automatic tumor segmentation. In this paper, we propose a counterfactual generation framework that not only achieves exceptional brain tumor segmentation performance without the need for pixel-level annotations, but also provides explainability. Our framework effectively separates class-related features from class-unrelated features of the samples, and generate new samples that preserve identity features while altering class attributes by embedding different class-related features. We perform topological data analysis on the extracted class-related features and obtain a globally explainable manifold, and for each abnormal sample to be segmented, a meaningful normal sample could be effectively generated with the guidance of the rule-based paths designed within the manifold for comparison for identifying the tumor regions. We evaluate our proposed method on two datasets, which demonstrates superior performance of brain tumor segmentation. The code is available at https://github.com/xrt11/tumor-segmentation.
Authors: Zidu Wang, Xiangyu Zhu, Jiang Yu, Tianshuo Zhang, Zhen Lei
Abstract: 3D textured face reconstruction from sketches applicable in many scenarios such as animation, 3D avatars, artistic design, missing people search, etc., is a highly promising but underdeveloped research topic. On the one hand, the stylistic diversity of sketches leads to existing sketch-to-3D-face methods only being able to handle pose-limited and realistically shaded sketches. On the other hand, texture plays a vital role in representing facial appearance, yet sketches lack this information, necessitating additional texture control in the reconstruction process. This paper proposes a novel method for reconstructing controllable textured and detailed 3D faces from sketches, named S2TD-Face. S2TD-Face introduces a two-stage geometry reconstruction framework that directly reconstructs detailed geometry from the input sketch. To keep geometry consistent with the delicate strokes of the sketch, we propose a novel sketch-to-geometry loss that ensures the reconstruction accurately fits the input features like dimples and wrinkles. Our training strategies do not rely on hard-to-obtain 3D face scanning data or labor-intensive hand-drawn sketches. Furthermore, S2TD-Face introduces a texture control module utilizing text prompts to select the most suitable textures from a library and seamlessly integrate them into the geometry, resulting in a 3D detailed face with controllable texture. S2TD-Face surpasses existing state-of-the-art methods in extensive quantitative and qualitative experiments. Our project is available at https://github.com/wang-zidu/S2TD-Face .
Authors: Muhammad Ahmad, Muhammad Hassaan Farooq Butt, Muhammad Usama, Hamad Ahmed Altuwaijri, Manual Mazzara, Salvatore Distenano
Abstract: Spatial-Spectral Mamba (SSM) improves computational efficiency and captures long-range dependencies, addressing Transformer limitations. However, traditional Mamba models overlook rich spectral information in HSIs and struggle with high dimensionality and sequential data. To address these issues, we propose the SSM with multi-head self-attention and token enhancement (MHSSMamba). This model integrates spectral and spatial information by enhancing spectral tokens and using multi-head attention to capture complex relationships between spectral bands and spatial locations. It also manages long-range dependencies and the sequential nature of HSI data, preserving contextual information across spectral bands. MHSSMamba achieved remarkable classification accuracies of 97.62\% on Pavia University, 96.92\% on the University of Houston, 96.85\% on Salinas, and 99.49\% on Wuhan-longKou datasets.
Authors: Simone Caldarella, Massimiliano Mancini, Elisa Ricci, Rahaf Aljundi
Abstract: Vision-Language Models (VLMs) combine visual and textual understanding, rendering them well-suited for diverse tasks like generating image captions and answering visual questions across various domains. However, these capabilities are built upon training on large amount of uncurated data crawled from the web. The latter may include sensitive information that VLMs could memorize and leak, raising significant privacy concerns. In this paper, we assess whether these vulnerabilities exist, focusing on identity leakage. Our study leads to three key findings: (i) VLMs leak identity information, even when the vision-language alignment and the fine-tuning use anonymized data; (ii) context has little influence on identity leakage; (iii) simple, widely used anonymization techniques, like blurring, are not sufficient to address the problem. These findings underscore the urgent need for robust privacy protection strategies when deploying VLMs. Ethical awareness and responsible development practices are essential to mitigate these risks.
Authors: Muhammad Ahmad, Muhammad Usama, Manual Mazzara
Abstract: Hyperspectral Imaging (HSI) has proven to be a powerful tool for capturing detailed spectral and spatial information across diverse applications. Despite the advancements in Deep Learning (DL) and Transformer architectures for HSI Classification (HSIC), challenges such as computational efficiency and the need for extensive labeled data persist. This paper introduces WaveMamba, a novel approach that integrates wavelet transformation with the Spatial-Spectral Mamba architecture to enhance HSIC. WaveMamba captures both local texture patterns and global contextual relationships in an end-to-end trainable model. The Wavelet-based enhanced features are then processed through the state-space architecture to model spatial-spectral relationships and temporal dependencies. The experimental results indicate that WaveMamba surpasses existing models, achieving an accuracy improvement of 4.5\% on the University of Houston dataset and a 2.0\% increase on the Pavia University dataset. These findings validate its effectiveness in addressing the complex data interactions inherent in HSIs.
Authors: Kushal Kumar Jain, Steve Grosz, Anoop M. Namboodiri, Anil K. Jain
Abstract: Forensic sketch-to-mugshot matching is a challenging task in face recognition, primarily hindered by the scarcity of annotated forensic sketches and the modality gap between sketches and photographs. To address this, we propose CLIP4Sketch, a novel approach that leverages diffusion models to generate a large and diverse set of sketch images, which helps in enhancing the performance of face recognition systems in sketch-to-mugshot matching. Our method utilizes Denoising Diffusion Probabilistic Models (DDPMs) to generate sketches with explicit control over identity and style. We combine CLIP and Adaface embeddings of a reference mugshot, along with textual descriptions of style, as the conditions to the diffusion model. We demonstrate the efficacy of our approach by generating a comprehensive dataset of sketches corresponding to mugshots and training a face recognition model on our synthetic data. Our results show significant improvements in sketch-to-mugshot matching accuracy over training on an existing, limited amount of real face sketch data, validating the potential of diffusion models in enhancing the performance of face recognition systems across modalities. We also compare our dataset with datasets generated using GAN-based methods to show its superiority.
Authors: Lutao Jiang, Hangyu Li, Lin Wang
Abstract: Text-to-3D content creation has recently received much attention, especially with the prevalence of 3D Gaussians Splatting. In general, GS-based methods comprise two key stages: initialization and rendering optimization. To achieve initialization, existing works directly apply random sphere initialization or 3D diffusion models, e.g., Point-E, to derive the initial shapes. However, such strategies suffer from two critical yet challenging problems: 1) the final shapes are still similar to the initial ones even after training; 2) shapes can be produced only from simple texts, e.g., "a dog", not for lexically richer texts, e.g., "a dog is sitting on the top of the airplane". To address these problems, this paper proposes a novel general framework to boost the 3D GS Initialization for text-to-3D generation upon the lexical richness. Our key idea is to aggregate 3D Gaussians into spatially uniform voxels to represent complex shapes while enabling the spatial interaction among the 3D Gaussians and semantic interaction between Gaussians and texts. Specifically, we first construct a voxelized representation, where each voxel holds a 3D Gaussian with its position, scale, and rotation fixed while setting opacity as the sole factor to determine a position's occupancy. We then design an initialization network mainly consisting of two novel components: 1) Global Information Perception (GIP) block and 2) Gaussians-Text Fusion (GTF) block. Such a design enables each 3D Gaussian to assimilate the spatial information from other areas and semantic information from texts. Extensive experiments show the superiority of our framework of high-quality 3D GS initialization against the existing methods, e.g., Shap-E, by taking lexically simple, medium, and hard texts. Also, our framework can be seamlessly plugged into SoTA training frameworks, e.g., LucidDreamer, for semantically consistent text-to-3D generation.
Authors: Wenbin Zou, Hongxia Gao, Weipeng Yang, Tongtong Liu
Abstract: Ultra-high-definition (UHD) technology has attracted widespread attention due to its exceptional visual quality, but it also poses new challenges for low-light image enhancement (LLIE) techniques. UHD images inherently possess high computational complexity, leading existing UHD LLIE methods to employ high-magnification downsampling to reduce computational costs, which in turn results in information loss. The wavelet transform not only allows downsampling without loss of information, but also separates the image content from the noise. It enables state space models (SSMs) to avoid being affected by noise when modeling long sequences, thus making full use of the long-sequence modeling capability of SSMs. On this basis, we propose Wave-Mamba, a novel approach based on two pivotal insights derived from the wavelet domain: 1) most of the content information of an image exists in the low-frequency component, less in the high-frequency component. 2) The high-frequency component exerts a minimal influence on the outcomes of low-light enhancement. Specifically, to efficiently model global content information on UHD images, we proposed a low-frequency state space block (LFSSBlock) by improving SSMs to focus on restoring the information of low-frequency sub-bands. Moreover, we propose a high-frequency enhance block (HFEBlock) for high-frequency sub-band information, which uses the enhanced low-frequency information to correct the high-frequency information and effectively restore the correct high-frequency details. Through comprehensive evaluation, our method has demonstrated superior performance, significantly outshining current leading techniques while maintaining a more streamlined architecture. The code is available at https://github.com/AlexZou14/Wave-Mamba.
Authors: Dong Huo, Zixin Guo, Xinxin Zuo, Zhihao Shi, Juwei Lu, Peng Dai, Songcen Xu, Li Cheng, Yee-Hong Yang
Abstract: Given a 3D mesh, we aim to synthesize 3D textures that correspond to arbitrary textual descriptions. Current methods for generating and assembling textures from sampled views often result in prominent seams or excessive smoothing. To tackle these issues, we present TexGen, a novel multi-view sampling and resampling framework for texture generation leveraging a pre-trained text-to-image diffusion model. For view consistent sampling, first of all we maintain a texture map in RGB space that is parameterized by the denoising step and updated after each sampling step of the diffusion model to progressively reduce the view discrepancy. An attention-guided multi-view sampling strategy is exploited to broadcast the appearance information across views. To preserve texture details, we develop a noise resampling technique that aids in the estimation of noise, generating inputs for subsequent denoising steps, as directed by the text prompt and current texture map. Through an extensive amount of qualitative and quantitative evaluations, we demonstrate that our proposed method produces significantly better texture quality for diverse 3D objects with a high degree of view consistency and rich appearance details, outperforming current state-of-the-art methods. Furthermore, our proposed texture generation technique can also be applied to texture editing while preserving the original identity. More experimental results are available at https://dong-huo.github.io/TexGen/
Authors: Muhammad Ali, Salman Khan
Abstract: The complex marine environment exacerbates the challenges of object detection manifold. Marine trash endangers the aquatic ecosystem, presenting a persistent challenge. Accurate detection of marine deposits is crucial for mitigating this harm. Our work addresses underwater object detection by enhancing image quality and evaluating detection methods. We use Detectron2's backbone with various base models and configurations for this task. We propose a novel channel stabilization technique alongside a simplified image enhancement model to reduce haze and color cast in training images, improving multi-scale object detection. Following image processing, we test different Detectron2 backbones for optimal detection accuracy. Additionally, we apply a sharpening filter with augmentation techniques to highlight object profiles for easier recognition. Results are demonstrated on the TrashCan Dataset, both instance and material versions. The best-performing backbone method incorporates our channel stabilization and augmentation techniques. We also compare our Detectron2 detection results with the Deformable Transformer. In the instance version of TrashCan 1.0, our method achieves a 9.53% absolute increase in average precision for small objects and a 7% absolute gain in bounding box detection compared to the baseline. The code will be available on Code: https://github.com/aliman80/Underwater- Object-Detection-via-Channel-Stablization
Authors: Danpei Zhao, Zhuoran Liu, Bo Yuan
Abstract: Improving search efficiency serves as one of the crucial objectives of Neural Architecture Search (NAS). However, many current approaches ignore the universality of the search strategy and fail to reduce the computational redundancy during the search process, especially in one-shot NAS architectures. Besides, current NAS methods show invalid reparameterization in non-linear search space, leading to poor efficiency in common search spaces like DARTS. In this paper, we propose TopoNAS, a model-agnostic approach for gradient-based one-shot NAS that significantly reduces searching time and memory usage by topological simplification of searchable paths. Firstly, we model the non-linearity in search spaces to reveal the parameterization difficulties. To improve the search efficiency, we present a topological simplification method and iteratively apply module-sharing strategies to simplify the topological structure of searchable paths. In addition, a kernel normalization technique is also proposed to preserve the search accuracy. Experimental results on the NASBench201 benchmark with various search spaces demonstrate the effectiveness of our method. It proves the proposed TopoNAS enhances the performance of various architectures in terms of search efficiency while maintaining a high level of accuracy. The project page is available at https://xdedss.github.io/topo_simplification.
Authors: Vito Mengers, Nicolas Roth, Oliver Brock, Klaus Obermayer, Martin Rolfs
Abstract: How we perceive objects around us depends on what we actively attend to, yet our eye movements depend on the perceived objects. Still, object segmentation and gaze behavior are typically treated as two independent processes. Drawing on an information processing pattern from robotics, we present a mechanistic model that simulates these processes for dynamic real-world scenes. Our image-computable model uses the current scene segmentation for object-based saccadic decision-making while using the foveated object to refine its scene segmentation recursively. To model this refinement, we use a Bayesian filter, which also provides an uncertainty estimate for the segmentation that we use to guide active scene exploration. We demonstrate that this model closely resembles observers' free viewing behavior, measured by scanpath statistics, including foveation duration and saccade amplitude distributions used for parameter fitting and higher-level statistics not used for fitting. These include how object detections, inspections, and returns are balanced and a delay of returning saccades without an explicit implementation of such temporal inhibition of return. Extensive simulations and ablation studies show that uncertainty promotes balanced exploration and that semantic object cues are crucial to form the perceptual units used in object-based attention. Moreover, we show how our model's modular design allows for extensions, such as incorporating saccadic momentum or pre-saccadic attention, to further align its output with human scanpaths.
Authors: Bingyu Li, Da Zhang, Zhiyuan Zhao, Junyu Gao, Xuelong Li
Abstract: Multimodal semantic segmentation shows significant potential for enhancing segmentation accuracy in complex scenes. However, current methods often incorporate specialized feature fusion modules tailored to specific modalities, thereby restricting input flexibility and increasing the number of training parameters. To address these challenges, we propose StitchFusion, a straightforward yet effective modal fusion framework that integrates large-scale pre-trained models directly as encoders and feature fusers. This approach facilitates comprehensive multi-modal and multi-scale feature fusion, accommodating any visual modal inputs. Specifically, Our framework achieves modal integration during encoding by sharing multi-modal visual information. To enhance information exchange across modalities, we introduce a multi-directional adapter module (MultiAdapter) to enable cross-modal information transfer during encoding. By leveraging MultiAdapter to propagate multi-scale information across pre-trained encoders during the encoding process, StitchFusion achieves multi-modal visual information integration during encoding. Extensive comparative experiments demonstrate that our model achieves state-of-the-art performance on four multi-modal segmentation datasets with minimal additional parameters. Furthermore, the experimental integration of MultiAdapter with existing Feature Fusion Modules (FFMs) highlights their complementary nature. Our code is available at StitchFusion_repo.
Authors: Peng Ding, Jingyu Wu, Jun Kuang, Dan Ma, Xuezhi Cao, Xunliang Cai, Shi Chen, Jiajun Chen, Shujian Huang
Abstract: Multi-modal Large Language Models (MLLMs) have demonstrated remarkable performance on various visual-language understanding and generation tasks. However, MLLMs occasionally generate content inconsistent with the given images, which is known as "hallucination". Prior works primarily center on evaluating hallucination using standard, unperturbed benchmarks, which overlook the prevalent occurrence of perturbed inputs in real-world scenarios-such as image cropping or blurring-that are critical for a comprehensive assessment of MLLMs' hallucination. In this paper, to bridge this gap, we propose Hallu-PI, the first benchmark designed to evaluate Hallucination in MLLMs within Perturbed Inputs. Specifically, Hallu-PI consists of seven perturbed scenarios, containing 1,260 perturbed images from 11 object types. Each image is accompanied by detailed annotations, which include fine-grained hallucination types, such as existence, attribute, and relation. We equip these annotations with a rich set of questions, making Hallu-PI suitable for both discriminative and generative tasks. Extensive experiments on 12 mainstream MLLMs, such as GPT-4V and Gemini-Pro Vision, demonstrate that these models exhibit significant hallucinations on Hallu-PI, which is not observed in unperturbed scenarios. Furthermore, our research reveals a severe bias in MLLMs' ability to handle different types of hallucinations. We also design two baselines specifically for perturbed scenarios, namely Perturbed-Reminder and Perturbed-ICL. We hope that our study will bring researchers' attention to the limitations of MLLMs when dealing with perturbed inputs, and spur further investigations to address this issue. Our code and datasets are publicly available at https://github.com/NJUNLP/Hallu-PI.
Authors: Yuanzhi Su, Siyuan Chen, Yuan-Gen Wang
Abstract: Class-incremental learning (CIL) thrives due to its success in processing the influx of information by learning from continuously added new classes while preventing catastrophic forgetting about the old ones. It is essential for the performance breakthrough of CIL to effectively refine past knowledge from the base model and balance it with new learning. However, such an issue has not yet been considered in current research. In this work, we explore the potential of CIL from these perspectives and propose a novel balanced residual distillation framework (BRD-CIL) to push the performance bar of CIL to a new higher level. Specifically, BRD-CIL designs a residual distillation learning strategy, which can dynamically expand the network structure to capture the residuals between the base and target models, effectively refining the past knowledge. Furthermore, BRD-CIL designs a balanced pseudo-label learning strategy by generating a guidance mask to reduce the preference for old classes, ensuring balanced learning from new and old classes. We apply the proposed BRD-CIL to a challenging 3D point cloud semantic segmentation task where the data are unordered and unstructured. Extensive experimental results demonstrate that BRD-CIL sets a new benchmark with an outstanding balance capability in class-biased scenarios.
Authors: Runze Yuan, Tao Liu, Zijia Dai, Yi-Fan Zuo, Laurent Kneip
Abstract: Event cameras are an interesting visual exteroceptive sensor that reacts to brightness changes rather than integrating absolute image intensities. Owing to this design, the sensor exhibits strong performance in situations of challenging dynamics and illumination conditions. While event-based simultaneous tracking and mapping remains a challenging problem, a number of recent works have pointed out the sensor's suitability for prior map-based tracking. By making use of cross-modal registration paradigms, the camera's ego-motion can be tracked across a large spectrum of illumination and dynamics conditions on top of accurate maps that have been created a priori by more traditional sensors. The present paper follows up on a recently introduced event-based geometric semi-dense tracking paradigm, and proposes the addition of inertial signals in order to robustify the estimation. More specifically, the added signals provide strong cues for pose initialization as well as regularization during windowed, multi-frame tracking. As a result, the proposed framework achieves increased performance under challenging illumination conditions as well as a reduction of the rate at which intermediate event representations need to be registered in order to maintain stable tracking across highly dynamic sequences. Our evaluation focuses on a diverse set of real world sequences and comprises a comparison of our proposed method against a purely event-based alternative running at different rates.
Authors: Muhammad Ahmad, Muhammad Hassaan Farooq Butt, Muhammad Usama, Adil Mehmood Khan, Manual Mazzara, Salvatore Distenano
Abstract: In recent years, Transformers have garnered significant attention for Hyperspectral Image Classification (HSIC) due to their self-attention mechanism, which provides strong classification performance. However, these models face major challenges in computational efficiency, as their complexity increases quadratically with the sequence length. The Mamba architecture, leveraging a State Space Model, offers a more efficient alternative to Transformers. This paper introduces the Spatial-Spectral Morphological Mamba (MorpMamba) model. In the MorpMamba model, a token generation module first converts the Hyperspectral Image (HSI) patch into spatial-spectral tokens. These tokens are then processed by a morphology block, which computes structural and shape information using depthwise separable convolutional operations. The extracted information is enhanced in a feature enhancement module that adjusts the spatial and spectral tokens based on the center region of the HSI sample, allowing for effective information fusion within each block. Subsequently, the tokens are refined in a multi-head self-attention block to further improve the feature space. Finally, the combined information is fed into the state space block for classification and the creation of the ground truth map. Experiments on widely used Hyperspectral (HS) datasets demonstrate that the MorpMamba model outperforms (parametric efficiency) both CNN and Transformer models.
Authors: Bohan Zhou, Jiangxing Wang, Zongqing Lu
Abstract: The in-context learning ability of Transformer models has brought new possibilities to visual navigation. In this paper, we focus on the video navigation setting, where an in-context navigation policy needs to be learned purely from videos in an offline manner, without access to the actual environment. For this setting, we propose Navigate Only Look Once (NOLO), a method for learning a navigation policy that possesses the in-context ability and adapts to new scenes by taking corresponding context videos as input without finetuning or re-training. To enable learning from videos, we first propose a pseudo action labeling procedure using optical flow to recover the action label from egocentric videos. Then, offline reinforcement learning is applied to learn the navigation policy. Through extensive experiments on different scenes, we show that our algorithm outperforms baselines by a large margin, which demonstrates the in-context learning ability of the learned policy.
Authors: Peng Zhang, Divya Chaudhary
Abstract: Cancer is a leading cause of death worldwide, necessitating advancements in early detection and treatment technologies. In this paper, we present a novel and highly efficient melanoma detection framework that synergistically combines the strengths of U-Net for segmentation and EfficientNet for the classification of skin images. The primary objective of our study is to enhance the accuracy and efficiency of melanoma detection through an innovative hybrid approach. We utilized the HAM10000 dataset to meticulously train the U-Net model, enabling it to precisely segment cancerous regions. Concurrently, we employed the ISIC 2020 dataset to train the EfficientNet model, optimizing it for the binary classification of skin cancer. Our hybrid model demonstrates a significant improvement in performance, achieving a remarkable accuracy of 99.01% on the ISIC 2020 dataset. This exceptional result underscores the superiority of our approach compared to existing model structures. By integrating the precise segmentation capabilities of U-Net with the advanced classification prowess of EfficientNet, our framework offers a comprehensive solution for melanoma detection. The results of our extensive experiments highlight the high accuracy and reliability of our method in both segmentation and classification tasks. This indicates the potential of our hybrid approach to significantly enhance cancer detection, providing a robust tool for medical professionals in the early diagnosis and treatment of melanoma. We believe that our framework can set a new benchmark in the field of automated skin cancer detection, encouraging further research and development in this crucial area of medical imaging.
Authors: Tamiris Abildayeva, Pakizar Shamoi
Abstract: Websites form the foundation of the Internet, serving as platforms for disseminating information and accessing digital resources. They allow users to engage with a wide range of content and services, enhancing the Internet's utility for all. The aesthetics of a website play a crucial role in its overall effectiveness and can significantly impact user experience, engagement, and satisfaction. This paper examines the importance of website design aesthetics in enhancing user experience, given the increasing number of internet users worldwide. It emphasizes the significant impact of first impressions, often formed within 50 milliseconds, on users' perceptions of a website's appeal and usability. We introduce a novel method for measuring website aesthetics based on color harmony and font popularity, using fuzzy logic to predict aesthetic preferences. We collected our own dataset, consisting of nearly 200 popular and frequently used website designs, to ensure relevance and adaptability to the dynamic nature of web design trends. Dominant colors from website screenshots were extracted using k-means clustering. The findings aim to improve understanding of the relationship between aesthetics and usability in website design.
Authors: Bin Han, Cleo Yau, Su Lei, Jonathan Gratch
Abstract: Emotion recognition in social situations is a complex task that requires integrating information from both facial expressions and the situational context. While traditional approaches to automatic emotion recognition have focused on decontextualized signals, recent research emphasizes the importance of context in shaping emotion perceptions. This paper contributes to the emerging field of context-based emotion recognition by leveraging psychological theories of human emotion perception to inform the design of automated methods. We propose an approach that combines emotion recognition methods with Bayesian Cue Integration (BCI) to integrate emotion inferences from decontextualized facial expressions and contextual knowledge inferred via Large-language Models. We test this approach in the context of interpreting facial expressions during a social task, the prisoner's dilemma. Our results provide clear support for BCI across a range of automatic emotion recognition methods. The best automated method achieved results comparable to human observers, suggesting the potential for this approach to advance the field of affective computing.
Authors: Gaia Fior, Carlo Fonda, Enrique Canessa
Abstract: The facilitation of STEM education can be enhanced by the provision of opportunities for learners to gain a better understanding of science through the utilization of tangible and visual examples. The objective of this work is to present an account of our experiences and activities carried out in Italian schools with this novel approach. The selection of projects and experiences discussed --in which students develop a range of core competencies such as collaboration, creativity, critical thinking, experimentation, prototyping, communication and problem-solving; include tangible complex 3D printed structures, large micro-controller board replicas and the visualization of wind dynamics and tiny invisible elementary particles among others. These hands-on experiences demonstrate the benefits on the use of digital fabrication technologies implemented within a FabLab for STEM learning.
Authors: Zichen Song, Jiakang Li, Songning Lai, Sitan Huang
Abstract: Spiking neural networks (SNNs) have shown promise in various dynamic visual tasks, yet those ready for practical deployment often lack the compactness and robustness essential in resource-limited and safety-critical settings. Prior research has predominantly concentrated on enhancing the compactness or robustness of artificial neural networks through strategies like network pruning and adversarial training, with little exploration into similar methodologies for SNNs. Robust pruning of SNNs aims to reduce computational overhead while preserving both accuracy and robustness. Current robust pruning approaches generally necessitate expert knowledge and iterative experimentation to establish suitable pruning criteria or auxiliary modules, thus constraining their broader application. Concurrently, evolutionary algorithms (EAs) have been employed to automate the pruning of artificial neural networks, delivering remarkable outcomes yet overlooking the aspect of robustness. In this work, we propose CCSRP, an innovative robust pruning method for SNNs, underpinned by cooperative co-evolution. Robust pruning is articulated as a tri-objective optimization challenge, striving to balance accuracy, robustness, and compactness concurrently, resolved through a cooperative co-evolutionary pruning framework that independently prunes filters across layers using EAs. Our experiments on CIFAR-10 and SVHN demonstrate that CCSRP can match or exceed the performance of the latest methodologies.
Authors: Omkar Kokane, Prabhat Sati, Mukul Lokhande, Santosh Kumar Vishvakarma
Abstract: This paper presents the Hybrid Overestimating Approximate Adder designed to enhance the performance in processing engines, specifically focused on edge AI applications. A novel Plus One Adder design is proposed as an incremental adder in the RCA chain, incorporating a Full Adder with an excess 1 alongside inputs A, B, and Cin. The design approximates outputs to 2 bit values to reduce hardware complexity and improve resource efficiency. The Plus One Adder is integrated into a dynamically reconfigurable HOAA, allowing runtime interchangeability between accurate and approximate overestimation modes. The proposed design is demonstrated for multiple applications, such as Twos complement subtraction and Rounding to even, and the Configurable Activation function, which are critical components of the Processing engine. Our approach shows 21 percent improvement in area efficiency and 33 percent reduction in power consumption, compared to state of the art designs with minimal accuracy loss. Thus, the proposed HOAA could be a promising solution for resource-constrained environments, offering ideal trade-offs between hardware efficiency vs computational accuracy.
Authors: Zhe Wang, Aladine Chetouani, Rachid Jennane, Yuhua Ru, Wasim Issa, Mohamed Jarraya
Abstract: Knee Osteoarthritis (KOA) is a common musculoskeletal disorder that significantly affects the mobility of older adults. In the medical domain, images containing temporal data are frequently utilized to study temporal dynamics and statistically monitor disease progression. While deep learning-based generative models for natural images have been widely researched, there are comparatively few methods available for synthesizing temporal knee X-rays. In this work, we introduce a novel deep-learning model designed to synthesize intermediate X-ray images between a specific patient's healthy knee and severe KOA stages. During the testing phase, based on a healthy knee X-ray, the proposed model can produce a continuous and effective sequence of KOA X-ray images with varying degrees of severity. Specifically, we introduce a Diffusion-based Morphing Model by modifying the Denoising Diffusion Probabilistic Model. Our approach integrates diffusion and morphing modules, enabling the model to capture spatial morphing details between source and target knee X-ray images and synthesize intermediate frames along a geodesic path. A hybrid loss consisting of diffusion loss, morphing loss, and supervision loss was employed. We demonstrate that our proposed approach achieves the highest temporal frame synthesis performance, effectively augmenting data for classification models and simulating the progression of KOA.
Authors: Caiwen Jiang, Xiaodan Xing, Zaixin Ou, Mianxin Liu, Walsh Simon, Guang Yang, Dinggang Shen
Abstract: The progression of Idiopathic Pulmonary Fibrosis (IPF) significantly correlates with higher patient mortality rates. Early detection of IPF progression is critical for initiating timely treatment, which can effectively slow down the advancement of the disease. However, the current clinical criteria define disease progression requiring two CT scans with a one-year interval, presenting a dilemma: a disease progression is identified only after the disease has already progressed. To this end, in this paper, we develop a novel diffusion model to accurately predict the progression of IPF by generating patient's follow-up CT scan from the initial CT scan. Specifically, from the clinical prior knowledge, we tailor improvements to the traditional diffusion model and propose a Clinically-Informed Residual Diffusion model, called CIResDiff. The key innovations of CIResDiff include 1) performing the target region pre-registration to align the lung regions of two CT scans at different time points for reducing the generation difficulty, 2) adopting the residual diffusion instead of traditional diffusion to enable the model focus more on differences (i.e., lesions) between the two CT scans rather than the largely identical anatomical content, and 3) designing the clinically-informed process based on CLIP technology to integrate lung function information which is highly relevant to diagnosis into the reverse process for assisting generation. Extensive experiments on clinical data demonstrate that our approach can outperform state-of-the-art methods and effectively predict the progression of IPF.
Authors: Caiwen Jiang, Tianyu Wang, Xiaodan Xing, Mianxin Liu, Guang Yang, Zhongxiang Ding, Dinggang Shen
Abstract: Ischemic stroke is a severe condition caused by the blockage of brain blood vessels, and can lead to the death of brain tissue due to oxygen deprivation. Thrombectomy has become a common treatment choice for ischemic stroke due to its immediate effectiveness. But, it carries the risk of postoperative cerebral hemorrhage. Clinically, multiple CT scans within 0-72 hours post-surgery are used to monitor for hemorrhage. However, this approach exposes radiation dose to patients, and may delay the detection of cerebral hemorrhage. To address this dilemma, we propose a novel prediction framework for measuring postoperative cerebral hemorrhage using only the patient's initial CT scan. Specifically, we introduce a dual-task mutual learning framework to takes the initial CT scan as input and simultaneously estimates both the follow-up CT scan and prognostic label to predict the occurrence of postoperative cerebral hemorrhage. Our proposed framework incorporates two attention mechanisms, i.e., self-attention and interactive attention. Specifically, the self-attention mechanism allows the model to focus more on high-density areas in the image, which are critical for diagnosis (i.e., potential hemorrhage areas). The interactive attention mechanism further models the dependencies between the interrelated generation and classification tasks, enabling both tasks to perform better than the case when conducted individually. Validated on clinical data, our method can generate follow-up CT scans better than state-of-the-art methods, and achieves an accuracy of 86.37% in predicting follow-up prognostic labels. Thus, our work thus contributes to the timely screening of post-thrombectomy cerebral hemorrhage, and could significantly reform the clinical process of thrombectomy and other similar operations related to stroke.
Authors: Chayan Banerjee, Kien Nguyen, Olivier Salvado, Truyen Tran, Clinton Fookes
Abstract: The incorporation of physical information in machine learning frameworks is transforming medical image analysis (MIA). By integrating fundamental knowledge and governing physical laws, these models achieve enhanced robustness and interpretability. In this work, we explore the utility of physics-informed approaches for MIA (PIMIA) tasks such as registration, generation, classification, and reconstruction. We present a systematic literature review of over 80 papers on physics-informed methods dedicated to MIA. We propose a unified taxonomy to investigate what physics knowledge and processes are modelled, how they are represented, and the strategies to incorporate them into MIA models. We delve deep into a wide range of image analysis tasks, from imaging, generation, prediction, inverse imaging (super-resolution and reconstruction), registration, and image analysis (segmentation and classification). For each task, we thoroughly examine and present in a tabular format the central physics-guided operation, the region of interest (with respect to human anatomy), the corresponding imaging modality, the dataset used for model training, the deep network architecture employed, and the primary physical process, equation, or principle utilized. Additionally, we also introduce a novel metric to compare the performance of PIMIA methods across different tasks and datasets. Based on this review, we summarize and distil our perspectives on the challenges, open research questions, and directions for future research. We highlight key open challenges in PIMIA, including selecting suitable physics priors and establishing a standardized benchmarking platform.
Authors: Kentaro Uno, Takehiro Matsuoka, Akiyoshi Uchida, Kazuya Yoshida
Abstract: With the boost in the number of spacecraft launches in the current decades, the space debris problem is daily becoming significantly crucial. For sustainable space utilization, the continuous removal of space debris is the most severe problem for humanity. To maximize the reliability of the debris capture mission in orbit, accurate motion estimation of the target is essential. Space debris has lost its attitude and orbit control capabilities, and its shape is unknown due to the break. This paper proposes the Structure from Motion-based algorithm to perform unknown shaped space debris motion estimation with limited resources, where only 2D images are required as input. The method then outputs the reconstructed shape of the unknown object and the relative pose trajectory between the target and the camera simultaneously, which are exploited to estimate the target's motion. The method is quantitatively validated with the realistic image dataset generated by the microgravity experiment in a 2D air-floating testbed and 3D kinematic simulation.
Authors: Seungeun Oh, Sihun Baek, Jihong Park, Hyelin Nam, Praneeth Vepakomma, Ramesh Raskar, Mehdi Bennis, Seong-Lyun Kim
Abstract: In computer vision, the vision transformer (ViT) has increasingly superseded the convolutional neural network (CNN) for improved accuracy and robustness. However, ViT's large model sizes and high sample complexity make it difficult to train on resource-constrained edge devices. Split learning (SL) emerges as a viable solution, leveraging server-side resources to train ViTs while utilizing private data from distributed devices. However, SL requires additional information exchange for weight updates between the device and the server, which can be exposed to various attacks on private training data. To mitigate the risk of data breaches in classification tasks, inspired from the CutMix regularization, we propose a novel privacy-preserving SL framework that injects Gaussian noise into smashed data and mixes randomly chosen patches of smashed data across clients, coined DP-CutMixSL. Our analysis demonstrates that DP-CutMixSL is a differentially private (DP) mechanism that strengthens privacy protection against membership inference attacks during forward propagation. Through simulations, we show that DP-CutMixSL improves privacy protection against membership inference attacks, reconstruction attacks, and label inference attacks, while also improving accuracy compared to DP-SL and DP-MixSL.
Authors: R\'ois\'in Luo, James McDermott, Colm O'Riordan
Abstract: Perturbation robustness evaluates the vulnerabilities of models, arising from a variety of perturbations, such as data corruptions and adversarial attacks. Understanding the mechanisms of perturbation robustness is critical for global interpretability. We present a model-agnostic, global mechanistic interpretability method to interpret the perturbation robustness of image models. This research is motivated by two key aspects. First, previous global interpretability works, in tandem with robustness benchmarks, e.g. mean corruption error (mCE), are not designed to directly interpret the mechanisms of perturbation robustness within image models. Second, we notice that the spectral signal-to-noise ratios (SNR) of perturbed natural images exponentially decay over the frequency. This power-law-like decay implies that: Low-frequency signals are generally more robust than high-frequency signals -- yet high classification accuracy can not be achieved by low-frequency signals alone. By applying Shapley value theory, our method axiomatically quantifies the predictive powers of robust features and non-robust features within an information theory framework. Our method, dubbed as \textbf{I-ASIDE} (\textbf{I}mage \textbf{A}xiomatic \textbf{S}pectral \textbf{I}mportance \textbf{D}ecomposition \textbf{E}xplanation), provides a unique insight into model robustness mechanisms. We conduct extensive experiments over a variety of vision models pre-trained on ImageNet to show that \textbf{I-ASIDE} can not only \textbf{measure} the perturbation robustness but also \textbf{provide interpretations} of its mechanisms.
Authors: Liuyuan Wen
Abstract: Generalized Zero-Shot Learning (GZSL) is a challenging task requiring accurate classification of both seen and unseen classes. Within this domain, Audio-visual GZSL emerges as an extremely exciting yet difficult task, given the inclusion of both visual and acoustic features as multi-modal inputs. Existing efforts in this field mostly utilize either embedding-based or generative-based methods. However, generative training is difficult and unstable, while embedding-based methods often encounter domain shift problem. Thus, we find it promising to integrate both methods into a unified framework to leverage their advantages while mitigating their respective disadvantages. Our study introduces a general framework employing out-of-distribution (OOD) detection, aiming to harness the strengths of both approaches. We first employ generative adversarial networks to synthesize unseen features, enabling the training of an OOD detector alongside classifiers for seen and unseen classes. This detector determines whether a test feature belongs to seen or unseen classes, followed by classification utilizing separate classifiers for each feature type. We test our framework on three popular audio-visual datasets and observe a significant improvement comparing to existing state-of-the-art works. Codes can be found in https://github.com/liuyuan-wen/AV-OOD-GZSL.
Authors: Yihao Ding, Jean Lee, Soyeon Caren Han
Abstract: Visually Rich Documents (VRDs) are essential in academia, finance, medical fields, and marketing due to their multimodal information content. Traditional methods for extracting information from VRDs depend on expert knowledge and manual labor, making them costly and inefficient. The advent of deep learning has revolutionized this process, introducing models that leverage multimodal information vision, text, and layout along with pretraining tasks to develop comprehensive document representations. These models have achieved state-of-the-art performance across various downstream tasks, significantly enhancing the efficiency and accuracy of information extraction from VRDs. In response to the growing demands and rapid developments in Visually Rich Document Understanding (VRDU), this paper provides a comprehensive review of deep learning-based VRDU frameworks. We systematically survey and analyze existing methods and benchmark datasets, categorizing them based on adopted strategies and downstream tasks. Furthermore, we compare different techniques used in VRDU models, focusing on feature representation and fusion, model architecture, and pretraining methods, while highlighting their strengths, limitations, and appropriate scenarios. Finally, we identify emerging trends and challenges in VRDU, offering insights into future research directions and practical applications. This survey aims to provide a thorough understanding of VRDU advancements, benefiting both academic and industrial sectors.
Authors: Xiaoshuang Li, Mingyuan Meng, Zimo Huang, Lei Bi, Eduardo Delamare, Dagan Feng, Bin Sheng, Jinman Kim
Abstract: Panoramic X-ray (PX) is a prevalent modality in dental practice for its wide availability and low cost. However, as a 2D projection image, PX does not contain 3D anatomical information, and therefore has limited use in dental applications that can benefit from 3D information, e.g., tooth angular misa-lignment detection and classification. Reconstructing 3D structures directly from 2D PX has recently been explored to address limitations with existing methods primarily reliant on Convolutional Neural Networks (CNNs) for direct 2D-to-3D mapping. These methods, however, are unable to correctly infer depth-axis spatial information. In addition, they are limited by the in-trinsic locality of convolution operations, as the convolution kernels only capture the information of immediate neighborhood pixels. In this study, we propose a progressive hybrid Multilayer Perceptron (MLP)-CNN pyra-mid network (3DPX) for 2D-to-3D oral PX reconstruction. We introduce a progressive reconstruction strategy, where 3D images are progressively re-constructed in the 3DPX with guidance imposed on the intermediate recon-struction result at each pyramid level. Further, motivated by the recent ad-vancement of MLPs that show promise in capturing fine-grained long-range dependency, our 3DPX integrates MLPs and CNNs to improve the semantic understanding during reconstruction. Extensive experiments on two large datasets involving 464 studies demonstrate that our 3DPX outperforms state-of-the-art 2D-to-3D oral reconstruction methods, including standalone MLP and transformers, in reconstruction quality, and also im-proves the performance of downstream angular misalignment classification tasks.
Authors: Xiaoshuai Chen, Wei Chen, Dongmyoung Lee, Yukun Ge, Nicolas Rojas, Petar Kormushev
Abstract: End-to-end robot learning, particularly for long-horizon tasks, often results in unpredictable outcomes and poor generalization. To address these challenges, we propose a novel Therblig-based Backbone Framework (TBBF) to enhance robot task understanding and transferability. This framework uses therbligs (basic action elements) as the backbone to decompose high-level robot tasks into elemental robot configurations, which are then integrated with current foundation models to improve task understanding. The approach consists of two stages: offline training and online testing. During the offline training stage, we developed the Meta-RGate SynerFusion (MGSF) network for accurate therblig segmentation across various tasks. In the online testing stage, after a one-shot demonstration of a new task is collected, our MGSF network extracts high-level knowledge, which is then encoded into the image using Action Registration (ActionREG). Additionally, the Large Language Model (LLM)-Alignment Policy for Visual Correction (LAP-VC) is employed to ensure precise action execution, facilitating trajectory transfer in novel robot scenarios. Experimental results validate these methods, achieving 94.37% recall in therblig segmentation and success rates of 94.4% and 80% in real-world online robot testing for simple and complex scenarios, respectively. Supplementary material is available at: https://sites.google.com/view/therbligsbasedbackbone/home
URLs: https://sites.google.com/view/therbligsbasedbackbone/home
Authors: Yue Duan, Zhangxuan Gu, Zhenzhe Ying, Lei Qi, Changhua Meng, Yinghuan Shi
Abstract: In the realm of cross-modal retrieval, seamlessly integrating diverse modalities within multimedia remains a formidable challenge, especially given the complexities introduced by noisy correspondence learning (NCL). Such noise often stems from mismatched data pairs, which is a significant obstacle distinct from traditional noisy labels. This paper introduces Pseudo-Classification based Pseudo-Captioning (PC$^2$) framework to address this challenge. PC$^2$ offers a threefold strategy: firstly, it establishes an auxiliary "pseudo-classification" task that interprets captions as categorical labels, steering the model to learn image-text semantic similarity through a non-contrastive mechanism. Secondly, unlike prevailing margin-based techniques, capitalizing on PC$^2$'s pseudo-classification capability, we generate pseudo-captions to provide more informative and tangible supervision for each mismatched pair. Thirdly, the oscillation of pseudo-classification is borrowed to assistant the correction of correspondence. In addition to technical contributions, we develop a realistic NCL dataset called Noise of Web (NoW), which could be a new powerful NCL benchmark where noise exists naturally. Empirical evaluations of PC$^2$ showcase marked improvements over existing state-of-the-art robust cross-modal retrieval techniques on both simulated and realistic datasets with various NCL settings. The contributed dataset and source code are released at https://github.com/alipay/PC2-NoiseofWeb.
Authors: Jheng-Hong Yang, Jimmy Lin
Abstract: Vision--Language Models (VLMs) have demonstrated success across diverse applications, yet their potential to assist in relevance judgments remains uncertain. This paper assesses the relevance estimation capabilities of VLMs, including CLIP, LLaVA, and GPT-4V, within a large-scale \textit{ad hoc} retrieval task tailored for multimedia content creation in a zero-shot fashion. Preliminary experiments reveal the following: (1) Both LLaVA and GPT-4V, encompassing open-source and closed-source visual-instruction-tuned Large Language Models (LLMs), achieve notable Kendall's $\tau \sim 0.4$ when compared to human relevance judgments, surpassing the CLIPScore metric. (2) While CLIPScore is strongly preferred, LLMs are less biased towards CLIP-based retrieval systems. (3) GPT-4V's score distribution aligns more closely with human judgments than other models, achieving a Cohen's $\kappa$ value of around 0.08, which outperforms CLIPScore at approximately -0.096. These findings underscore the potential of LLM-powered VLMs in enhancing relevance judgments.
Authors: Ruoxuan Feng, Di Hu, Wenke Ma, Xuelong Li
Abstract: Humans possess a remarkable talent for flexibly alternating to different senses when interacting with the environment. Picture a chef skillfully gauging the timing of ingredient additions and controlling the heat according to the colors, sounds, and aromas, seamlessly navigating through every stage of the complex cooking process. This ability is founded upon a thorough comprehension of task stages, as achieving the sub-goal within each stage can necessitate the utilization of different senses. In order to endow robots with similar ability, we incorporate the task stages divided by sub-goals into the imitation learning process to accordingly guide dynamic multi-sensory fusion. We propose MS-Bot, a stage-guided dynamic multi-sensory fusion method with coarse-to-fine stage understanding, which dynamically adjusts the priority of modalities based on the fine-grained state within the predicted current stage. We train a robot system equipped with visual, auditory, and tactile sensors to accomplish challenging robotic manipulation tasks: pouring and peg insertion with keyway. Experimental results indicate that our approach enables more effective and explainable dynamic fusion, aligning more closely with the human fusion process than existing methods.
Authors: Yilun Hua, Yoav Artzi
Abstract: Humans spontaneously use increasingly efficient language as interactions progress, by adapting and forming ad-hoc conventions. This phenomenon has been studied extensively using reference games, showing properties of human language that go beyond relaying intents. It remains unexplored whether multimodal large language models (MLLMs) similarly increase communication efficiency during interactions, and what mechanisms they may adopt for this purpose. We introduce ICCA, an automated framework to evaluate such conversational adaptation as an in-context behavior in MLLMs. We evaluate several state-of-the-art MLLMs, and observe that while they may understand the increasingly efficient language of their interlocutor, they do not spontaneously make their own language more efficient over time. This latter ability can only be elicited in some models (e.g., GPT-4) with heavy-handed prompting. This shows that this property of linguistic interaction does not arise from current training regimes, even though it is a common hallmark of human language. ICCA is available at https://github.com/lil-lab/ICCA.
Authors: Li Zhang, Jiachen Lu, Sixiao Zheng, Xinxuan Zhao, Xiatian Zhu, Yanwei Fu, Tao Xiang, Jianfeng Feng, Philip H. S. Torr
Abstract: The emergence of vision transformers (ViTs) in image classification has shifted the methodologies for visual representation learning. In particular, ViTs learn visual representation at full receptive field per layer across all the image patches, in comparison to the increasing receptive fields of CNNs across layers and other alternatives (e.g., large kernels and atrous convolution). In this work, for the first time we explore the global context learning potentials of ViTs for dense visual prediction (e.g., semantic segmentation). Our motivation is that through learning global context at full receptive field layer by layer, ViTs may capture stronger long-range dependency information, critical for dense prediction tasks. We first demonstrate that encoding an image as a sequence of patches, a vanilla ViT without local convolution and resolution reduction can yield stronger visual representation for semantic segmentation. For example, our model, termed as SEgmentation TRansformer (SETR), excels on ADE20K (50.28% mIoU, the first position in the test leaderboard on the day of submission) and performs competitively on Cityscapes. However, the basic ViT architecture falls short in broader dense prediction applications, such as object detection and instance segmentation, due to its lack of a pyramidal structure, high computational demand, and insufficient local context. For tackling general dense visual prediction tasks in a cost-effective manner, we further formulate a family of Hierarchical Local-Global (HLG) Transformers, characterized by local attention within windows and global-attention across windows in a pyramidal architecture. Extensive experiments show that our methods achieve appealing performance on a variety of dense prediction tasks (e.g., object detection and instance segmentation and semantic segmentation) as well as image classification.
Authors: Arun V. Reddy, Ketul Shah, William Paul, Rohita Mocharla, Judy Hoffman, Kapil D. Katyal, Dinesh Manocha, Celso M. de Melo, Rama Chellappa
Abstract: Human action recognition is a challenging problem, particularly when there is high variability in factors such as subject appearance, backgrounds and viewpoint. While deep neural networks (DNNs) have been shown to perform well on action recognition tasks, they typically require large amounts of high-quality labeled data to achieve robust performance across a variety of conditions. Synthetic data has shown promise as a way to avoid the substantial costs and potential ethical concerns associated with collecting and labeling enormous amounts of data in the real-world. However, synthetic data may differ from real data in important ways. This phenomenon, known as \textit{domain shift}, can limit the utility of synthetic data in robotics applications. To mitigate the effects of domain shift, substantial effort is being dedicated to the development of domain adaptation (DA) techniques. Yet, much remains to be understood about how best to develop these techniques. In this paper, we introduce a new dataset called Robot Control Gestures (RoCoG-v2). The dataset is composed of both real and synthetic videos from seven gesture classes, and is intended to support the study of synthetic-to-real domain shift for video-based action recognition. Our work expands upon existing datasets by focusing the action classes on gestures for human-robot teaming, as well as by enabling investigation of domain shift in both ground and aerial views. We present baseline results using state-of-the-art action recognition and domain adaptation algorithms and offer initial insight on tackling the synthetic-to-real and ground-to-air domain shifts.
Authors: Rui Zhang, Luziwei Leng, Kaiwei Che, Hu Zhang, Jie Cheng, Qinghai Guo, Jiangxing Liao, Ran Cheng
Abstract: Spiking neural networks (SNNs), known for their low-power, event-driven computation and intrinsic temporal dynamics, are emerging as promising solutions for processing dynamic, asynchronous signals from event-based sensors. Despite their potential, SNNs face challenges in training and architectural design, resulting in limited performance in challenging event-based dense prediction tasks compared to artificial neural networks (ANNs). In this work, we develop an efficient spiking encoder-decoder network (SpikingEDN) for large-scale event-based semantic segmentation tasks. To enhance the learning efficiency from dynamic event streams, we harness the adaptive threshold which improves network accuracy, sparsity and robustness in streaming inference. Moreover, we develop a dual-path Spiking Spatially-Adaptive Modulation module, which is specifically tailored to enhance the representation of sparse events and multi-modal inputs, thereby considerably improving network performance. Our SpikingEDN attains a mean intersection over union (MIoU) of 72.57\% on the DDD17 dataset and 58.32\% on the larger DSEC-Semantic dataset, showing competitive results to the state-of-the-art ANNs while requiring substantially fewer computational resources. Our results shed light on the untapped potential of SNNs in event-based vision applications. The source code will be made publicly available.
Authors: Giang Nguyen, Valerie Chen, Mohammad Reza Taesiri, Anh Totti Nguyen
Abstract: Nearest neighbors (NN) are traditionally used to compute final decisions, e.g., in Support Vector Machines or k-NN classifiers, and to provide users with explanations for the model's decision. In this paper, we show a novel utility of nearest neighbors: To improve predictions of a frozen, pretrained classifier C. We leverage an image comparator S that (1) compares the input image with NN images from the top-K most probable classes; and (2) uses S' output scores to weight the confidence scores of C. Our method consistently improves fine-grained image classification accuracy on CUB-200, Cars-196, and Dogs-120. Also, a human study finds that showing lay users our probable-class nearest neighbors (PCNN) reduces over-reliance on AI, thus improving their decision accuracy over prior work which only shows only the top-1 class examples.
Authors: Zhuokai Zhao, Harish Palani, Tianyi Liu, Lena Evans, Ruth Toner
Abstract: Multimodal deep learning, especially vision-language models, have gained significant traction in recent years, greatly improving performance on many downstream tasks, including content moderation and violence detection. However, standard multimodal approaches often assume consistent modalities between training and inference, limiting applications in many real-world use cases, as some modalities may not be available during inference. While existing research mitigates this problem through reconstructing the missing modalities, they unavoidably increase unnecessary computational cost, which could be just as critical, especially for large, deployed infrastructures in industry. To this end, we propose a novel guidance network that promotes knowledge sharing during training, taking advantage of the multimodal representations to train better single-modality models to be used for inference. Real-world experiments in violence detection shows that our proposed framework trains single-modality models that significantly outperform traditionally trained counterparts, while avoiding increases in computational cost for inference.
Authors: Amr Gomaa, Robin Zitt, Guillermo Reyes, Antonio Kr\"uger
Abstract: Creating a diverse and comprehensive dataset of hand gestures for dynamic human-machine interfaces in the automotive domain can be challenging and time-consuming. To overcome this challenge, we propose using synthetic gesture datasets generated by virtual 3D models. Our framework utilizes Unreal Engine to synthesize realistic hand gestures, offering customization options and reducing the risk of overfitting. Multiple variants, including gesture speed, performance, and hand shape, are generated to improve generalizability. In addition, we simulate different camera locations and types, such as RGB, infrared, and depth cameras, without incurring additional time and cost to obtain these cameras. Experimental results demonstrate that our proposed framework, SynthoGestures (https://github.com/amrgomaaelhady/SynthoGestures), improves gesture recognition accuracy and can replace or augment real-hand datasets. By saving time and effort in the creation of the data set, our tool accelerates the development of gesture recognition systems for automotive applications.
Authors: Simone Caldarella, Elisa Ricci, Rahaf Aljundi
Abstract: Object-based Novelty Detection (ND) aims to identify unknown objects that do not belong to classes seen during training by an object detection model. The task is particularly crucial in real-world applications, as it allows to avoid potentially harmful behaviours, e.g. as in the case of object detection models adopted in a self-driving car or in an autonomous robot. Traditional approaches to ND focus on one time offline post processing of the pretrained object detection output, leaving no possibility to improve the model robustness after training and discarding the abundant amount of out-of-distribution data encountered during deployment. In this work, we propose a novel framework for object-based ND, assuming that human feedback can be requested on the predicted output and later incorporated to refine the ND model without negatively affecting the main object detection performance. This refinement operation is repeated whenever new feedback is available. To tackle this new formulation of the problem for object detection, we propose a lightweight ND module attached on top of a pre-trained object detection model, which is incrementally updated through a feedback loop. We also propose a new benchmark to evaluate methods on this new setting and test extensively our ND approach against baselines, showing increased robustness and a successful incorporation of the received feedback.
Authors: Somayeh Hussaini, Michael Milford, Tobias Fischer
Abstract: In robotics, Spiking Neural Networks (SNNs) are increasingly recognized for their largely-unrealized potential energy efficiency and low latency particularly when implemented on neuromorphic hardware. Our paper highlights three advancements for SNNs in Visual Place Recognition (VPR). Firstly, we propose Modular SNNs, where each SNN represents a set of non-overlapping geographically distinct places, enabling scalable networks for large environments. Secondly, we present Ensembles of Modular SNNs, where multiple networks represent the same place, significantly enhancing accuracy compared to single-network models. Each of our Modular SNN modules is compact, comprising only 1500 neurons and 474k synapses, making them ideally suited for ensembling due to their small size. Lastly, we investigate the role of sequence matching in SNN-based VPR, a technique where consecutive images are used to refine place recognition. We analyze the responsiveness of SNNs to ensembling and sequence matching compared to other VPR techniques. Our contributions highlight the viability of SNNs for VPR, offering scalable and robust solutions, and paving the way for their application in various energy-sensitive robotic tasks.
Authors: Gustav Bredell, Marcel Fischer, Przemyslaw Szostak, Samaneh Abbasi-Sureshjani, Alvaro Gomariz
Abstract: Digital pathology has significantly advanced disease detection and pathologist efficiency through the analysis of gigapixel whole-slide images (WSI). In this process, WSIs are first divided into patches, for which a feature extractor model is applied to obtain feature vectors, which are subsequently processed by an aggregation model to predict the respective WSI label. With the rapid evolution of representation learning, numerous new feature extractor models, often termed foundational models, have emerged. Traditional evaluation methods rely on a static downstream aggregation model setup, encompassing a fixed architecture and hyperparameters, a practice we identify as potentially biasing the results. Our study uncovers a sensitivity of feature extractor models towards aggregation model configurations, indicating that performance comparability can be skewed based on the chosen configurations. By accounting for this sensitivity, we find that the performance of many current feature extractor models is notably similar. We support this insight by evaluating seven feature extractor models across three different datasets with 162 different aggregation model configurations. This comprehensive approach provides a more nuanced understanding of the feature extractors' sensitivity to various aggregation model configurations, leading to a fairer and more accurate assessment of new foundation models in digital pathology.
Authors: Haoyu Guo, He Zhu, Sida Peng, Yuang Wang, Yujun Shen, Ruizhen Hu, Xiaowei Zhou
Abstract: This paper addresses the challenge of 3D instance segmentation by simultaneously leveraging 3D geometric and multi-view image information. Many previous works have applied deep learning techniques to 3D point clouds for instance segmentation. However, these methods often failed to generalize to various types of scenes due to the scarcity and low-diversity of labeled 3D point cloud data. Some recent works have attempted to lift 2D instance segmentations to 3D within a bottom-up framework. The inconsistency in 2D instance segmentations among views can substantially degrade the performance of 3D segmentation. In this work, we introduce a novel 3D-to-2D query framework to effectively exploit 2D segmentation models for 3D instance segmentation. Specifically, we pre-segment the scene into several superpoints in 3D, formulating the task into a graph cut problem. The superpoint graph is constructed based on 2D segmentation models, where node features are obtained from multi-view image features and edge weights are computed based on multi-view segmentation results, enabling the better generalization ability. To process the graph, we train a graph neural network using pseudo 3D labels from 2D segmentation models. Experimental results on the ScanNet, ScanNet++ and KITTI-360 datasets demonstrate that our method achieves robust segmentation performance and can generalize across different types of scenes. Our project page is available at https://zju3dv.github.io/sam_graph.
Authors: Ziying Song, Lin Liu, Feiyang Jia, Yadan Luo, Guoxin Zhang, Lei Yang, Li Wang, Caiyan Jia
Abstract: In the realm of modern autonomous driving, the perception system is indispensable for accurately assessing the state of the surrounding environment, thereby enabling informed prediction and planning. The key step to this system is related to 3D object detection that utilizes vehicle-mounted sensors such as LiDAR and cameras to identify the size, the category, and the location of nearby objects. Despite the surge in 3D object detection methods aimed at enhancing detection precision and efficiency, there is a gap in the literature that systematically examines their resilience against environmental variations, noise, and weather changes. This study emphasizes the importance of robustness, alongside accuracy and latency, in evaluating perception systems under practical scenarios. Our work presents an extensive survey of camera-only, LiDAR-only, and multi-modal 3D object detection algorithms, thoroughly evaluating their trade-off between accuracy, latency, and robustness, particularly on datasets like KITTI-C and nuScenes-C to ensure fair comparisons. Among these, multi-modal 3D detection approaches exhibit superior robustness, and a novel taxonomy is introduced to reorganize the literature for enhanced clarity. This survey aims to offer a more practical perspective on the current capabilities and the constraints of 3D object detection algorithms in real-world applications, thus steering future research towards robustness-centric advancements.
Authors: Kibum Kim, Kanghoon Yoon, Yeonjun In, Jinyoung Moon, Donghyun Kim, Chanyoung Park
Abstract: Scene graph generation (SGG) models have suffered from inherent problems regarding the benchmark datasets such as the long-tailed predicate distribution and missing annotation problems. In this work, we aim to alleviate the long-tailed problem of SGG by utilizing unannotated triplets. To this end, we introduce a Self-Training framework for SGG (ST-SGG) that assigns pseudo-labels for unannotated triplets based on which the SGG models are trained. While there has been significant progress in self-training for image recognition, designing a self-training framework for the SGG task is more challenging due to its inherent nature such as the semantic ambiguity and the long-tailed distribution of predicate classes. Hence, we propose a novel pseudo-labeling technique for SGG, called Class-specific Adaptive Thresholding with Momentum (CATM), which is a model-agnostic framework that can be applied to any existing SGG models. Furthermore, we devise a graph structure learner (GSL) that is beneficial when adopting our proposed self-training framework to the state-of-the-art message-passing neural network (MPNN)-based SGG models. Our extensive experiments verify the effectiveness of ST-SGG on various SGG models, particularly in enhancing the performance on fine-grained predicate classes.
Authors: Adriano D'Alessandro, Ali Mahdavi-Amiri, Ghassan Hamarneh
Abstract: Object counting methods typically rely on manually annotated datasets. The cost of creating such datasets has restricted the versatility of these networks to count objects from specific classes (such as humans or penguins), and counting objects from diverse categories remains a challenge. The availability of robust text-to-image latent diffusion models (LDMs) raises the question of whether these models can be utilized to generate counting datasets. However, LDMs struggle to create images with an exact number of objects based solely on text prompts but they can be used to offer a dependable \textit{sorting} signal by adding and removing objects within an image. Leveraging this data, we initially introduce an unsupervised sorting methodology to learn object-related features that are subsequently refined and anchored for counting purposes using counting data generated by LDMs. Further, we present a density classifier-guided method for dividing an image into patches containing objects that can be reliably counted. Consequently, we can generate counting data for any type of object and count them in an unsupervised manner. Our approach outperforms other unsupervised and few-shot alternatives and is not restricted to specific object classes for which counting data is available. Code to be released upon acceptance.
Authors: Opher Bar Nathan, Deborah Levy, Tali Treibitz, Dan Rosenbaum
Abstract: Underwater image restoration is a challenging task because of water effects that increase dramatically with distance. This is worsened by lack of ground truth data of clean scenes without water. Diffusion priors have emerged as strong image restoration priors. However, they are often trained with a dataset of the desired restored output, which is not available in our case. We also observe that using only color data is insufficient, and therefore augment the prior with a depth channel. We train an unconditional diffusion model prior on the joint space of color and depth, using standard RGBD datasets of natural outdoor scenes in air. Using this prior together with a novel guidance method based on the underwater image formation model, we generate posterior samples of clean images, removing the water effects. Even though our prior did not see any underwater images during training, our method outperforms state-of-the-art baselines for image restoration on very challenging scenes. Our code, models and data are available on the project website.
Authors: Zhengyi Zhao, Chen Song, Xiaodong Gu, Yuan Dong, Qi Zuo, Weihao Yuan, Liefeng Bo, Zilong Dong, Qixing Huang
Abstract: A fundamental problem in the texturing of 3D meshes using pre-trained text-to-image models is to ensure multi-view consistency. State-of-the-art approaches typically use diffusion models to aggregate multi-view inputs, where common issues are the blurriness caused by the averaging operation in the aggregation step or inconsistencies in local features. This paper introduces an optimization framework that proceeds in four stages to achieve multi-view consistency. Specifically, the first stage generates an over-complete set of 2D textures from a predefined set of viewpoints using an MV-consistent diffusion process. The second stage selects a subset of views that are mutually consistent while covering the underlying 3D model. We show how to achieve this goal by solving semi-definite programs. The third stage performs non-rigid alignment to align the selected views across overlapping regions. The fourth stage solves an MRF problem to associate each mesh face with a selected view. In particular, the third and fourth stages are iterated, with the cuts obtained in the fourth stage encouraging non-rigid alignment in the third stage to focus on regions close to the cuts. Experimental results show that our approach significantly outperforms baseline approaches both qualitatively and quantitatively. Project page: https://aigc3d.github.io/ConsistenTex.
Authors: Jianqiang Xiao, Weiwen Guo, Junfeng Liu, Mengze Li
Abstract: In the field of computer vision, data augmentation is widely used to enrich the feature complexity of training datasets with deep learning techniques. However, regarding the generalization capabilities of models, the difference in artificial features generated by data augmentation and natural visual features has not been fully revealed. This study introduces the concept of "visual representation variables" to define the possible visual variations in a task as a joint distribution of these variables. We focus on the visual representation variable "illumination", by simulating its distribution degradation and examining how data augmentation techniques enhance model performance on a classification task. Our goal is to investigate the differences in generalization between models trained with augmented data and those trained under real-world illumination conditions. Results indicate that after applying various data augmentation methods, model performance has significantly improved. Yet, a noticeable generalization gap still exists after utilizing various data augmentation methods, emphasizing the critical role of feature diversity in the training set for enhancing model generalization.
Authors: Mykola Lavreniuk
Abstract: Self-supervised monocular depth estimation has garnered considerable attention for its applications in autonomous driving and robotics. While recent methods have made strides in leveraging techniques like the Self Query Layer (SQL) to infer depth from motion, they often overlook the potential of strengthening pose information. In this paper, we introduce SPIdepth, a novel approach that prioritizes enhancing the pose network for improved depth estimation. Building upon the foundation laid by SQL, SPIdepth emphasizes the importance of pose information in capturing fine-grained scene structures. By enhancing the pose network's capabilities, SPIdepth achieves remarkable advancements in scene understanding and depth estimation. Experimental results on benchmark datasets such as KITTI, Cityscapes, and Make3D showcase SPIdepth's state-of-the-art performance, surpassing previous methods by significant margins. Specifically, SPIdepth tops the self-supervised KITTI benchmark. Additionally, SPIdepth achieves the lowest AbsRel (0.029), SqRel (0.069), and RMSE (1.394) on KITTI, establishing new state-of-the-art results. On Cityscapes, SPIdepth shows improvements over SQLdepth of 21.7% in AbsRel, 36.8% in SqRel, and 16.5% in RMSE, even without using motion masks. On Make3D, SPIdepth in zero-shot outperforms all other models. Remarkably, SPIdepth achieves these results using only a single image for inference, surpassing even methods that utilize video sequences for inference, thus demonstrating its efficacy and efficiency in real-world applications. Our approach represents a significant leap forward in self-supervised monocular depth estimation, underscoring the importance of strengthening pose information for advancing scene understanding in real-world applications.
Authors: Zhilin Huang, Yijie Yu, Ling Yang, Chujun Qin, Bing Zheng, Xiawu Zheng, Zikun Zhou, Yaowei Wang, Wenming Yang
Abstract: With the advancement of AIGC, video frame interpolation (VFI) has become a crucial component in existing video generation frameworks, attracting widespread research interest. For the VFI task, the motion estimation between neighboring frames plays a crucial role in avoiding motion ambiguity. However, existing VFI methods always struggle to accurately predict the motion information between consecutive frames, and this imprecise estimation leads to blurred and visually incoherent interpolated frames. In this paper, we propose a novel diffusion framework, motion-aware latent diffusion models (MADiff), which is specifically designed for the VFI task. By incorporating motion priors between the conditional neighboring frames with the target interpolated frame predicted throughout the diffusion sampling procedure, MADiff progressively refines the intermediate outcomes, culminating in generating both visually smooth and realistic results. Extensive experiments conducted on benchmark datasets demonstrate that our method achieves state-of-the-art performance significantly outperforming existing approaches, especially under challenging scenarios involving dynamic textures with complex motion.
Authors: Heitor R. Medeiros, David Latortue, Eric Granger, Marco Pedersoli
Abstract: In real-world scenarios, using multiple modalities like visible (RGB) and infrared (IR) can greatly improve the performance of a predictive task such as object detection (OD). Multimodal learning is a common way to leverage these modalities, where multiple modality-specific encoders and a fusion module are used to improve performance. In this paper, we tackle a different way to employ RGB and IR modalities, where only one modality or the other is observed by a single shared vision encoder. This realistic setting requires a lower memory footprint and is more suitable for applications such as autonomous driving and surveillance, which commonly rely on RGB and IR data. However, when learning a single encoder on multiple modalities, one modality can dominate the other, producing uneven recognition results. This work investigates how to efficiently leverage RGB and IR modalities to train a common transformer-based OD vision encoder, while countering the effects of modality imbalance. For this, we introduce a novel training technique to Mix Patches (MiPa) from the two modalities, in conjunction with a patch-wise modality agnostic module, for learning a common representation of both modalities. Our experiments show that MiPa can learn a representation to reach competitive results on traditional RGB/IR benchmarks while only requiring a single modality during inference. Our code is available at: https://github.com/heitorrapela/MiPa.
Authors: Kaiyu Li, Xiangyong Cao, Yupeng Deng, Junmin Liu, Deyu Meng, Zhi Wang
Abstract: Change Detection (CD) aims to identify pixels with semantic changes between images. However, annotating massive numbers of pixel-level images is labor-intensive and costly, especially for multi-temporal images, which require pixel-wise comparisons by human experts. Considering the excellent performance of visual language models (VLMs) for zero-shot, open-vocabulary, etc. with prompt-based reasoning, it is promising to utilize VLMs to make better CD under limited labeled data. In this paper, we propose a VLM guidance-based semi-supervised CD method, namely SemiCD-VL. The insight of SemiCD-VL is to synthesize free change labels using VLMs to provide additional supervision signals for unlabeled data. However, almost all current VLMs are designed for single-temporal images and cannot be directly applied to bi- or multi-temporal images. Motivated by this, we first propose a VLM-based mixed change event generation (CEG) strategy to yield pseudo labels for unlabeled CD data. Since the additional supervised signals provided by these VLM-driven pseudo labels may conflict with the pseudo labels from the consistency regularization paradigm (e.g. FixMatch), we propose the dual projection head for de-entangling different signal sources. Further, we explicitly decouple the bi-temporal images semantic representation through two auxiliary segmentation decoders, which are also guided by VLM. Finally, to make the model more adequately capture change representations, we introduce metric-aware supervision by feature-level contrastive loss in auxiliary branches. Extensive experiments show the advantage of SemiCD-VL. For instance, SemiCD-VL improves the FixMatch baseline by +5.3 IoU on WHU-CD and by +2.4 IoU on LEVIR-CD with 5% labels. In addition, our CEG strategy, in an un-supervised manner, can achieve performance far superior to state-of-the-art un-supervised CD methods.
Authors: Yibo Liu, Jinjun Shan, Amaldev Haridevan, Shuo Zhang
Abstract: Point cloud registration is a prerequisite for many applications in computer vision and robotics. Most existing methods focus on pairwise registration of two point clouds with high overlap. Although there have been some methods for low overlap cases, they struggle in degraded scenarios. This paper introduces a novel framework dubbed L-PR, designed to register unordered low overlap multiview point clouds leveraging LiDAR fiducial markers. We refer to them as LiDAR fiducial markers, but they are the same as the popular AprilTag and ArUco markers, thin sheets of paper that do not affect the 3D geometry of the environment. We first propose an improved adaptive threshold marker detection method to provide robust detection results when the viewpoints among point clouds change dramatically. Then, we formulate the unordered multiview point cloud registration problem as a maximum a-posteriori (MAP) problem and develop a framework consisting of two levels of graphs to address it. The first-level graph, constructed as a weighted graph, is designed to efficiently and optimally infer initial values of scan poses from the unordered set. The second-level graph is constructed as a factor graph. By globally optimizing the variables on the graph, including scan poses, marker poses, and marker corner positions, we tackle the MAP problem. We conduct both qualitative and quantitative experiments to demonstrate that the proposed method surpasses previous state-of-the-art (SOTA) methods and to showcase that L-PR can serve as a low-cost and efficient tool for 3D asset collection and training data collection. In particular, we collect a new dataset named Livox-3DMatch using L-PR and incorporate it into the training of the SOTA learning-based method, SGHR, which brings evident improvements for SGHR on various benchmarks.
Authors: Reyhane Askari Hemmat, Melissa Hall, Alicia Sun, Candace Ross, Michal Drozdzal, Adriana Romero-Soriano
Abstract: With the growing popularity of text-to-image generative models, there has been increasing focus on understanding their risks and biases. Recent work has found that state-of-the-art models struggle to depict everyday objects with the true diversity of the real world and have notable gaps between geographic regions. In this work, we aim to increase the diversity of generated images of common objects such that per-region variations are representative of the real world. We introduce an inference time intervention, contextualized Vendi Score Guidance (c-VSG), that guides the backwards steps of latent diffusion models to increase the diversity of a sample as compared to a "memory bank" of previously generated images while constraining the amount of variation within that of an exemplar set of real-world contextualizing images. We evaluate c-VSG with two geographically representative datasets and find that it substantially increases the diversity of generated images, both for the worst performing regions and on average, while simultaneously maintaining or improving image quality and consistency. Additionally, qualitative analyses reveal that diversity of generated images is significantly improved, including along the lines of reductive region portrayals present in the original model. We hope that this work is a step towards text-to-image generative models that reflect the true geographic diversity of the world.
Authors: Se Jin Park, Chae Won Kim, Hyeongseop Rha, Minsu Kim, Joanna Hong, Jeong Hun Yeo, Yong Man Ro
Abstract: In this paper, we introduce a novel Face-to-Face spoken dialogue model. It processes audio-visual speech from user input and generates audio-visual speech as the response, marking the initial step towards creating an avatar chatbot system without relying on intermediate text. To this end, we newly introduce MultiDialog, the first large-scale multimodal (i.e., audio and visual) spoken dialogue corpus containing 340 hours of approximately 9,000 dialogues, recorded based on the open domain dialogue dataset, TopicalChat. The MultiDialog contains parallel audio-visual recordings of conversation partners acting according to the given script with emotion annotations, which we expect to open up research opportunities in multimodal synthesis. Our Face-to-Face spoken dialogue model incorporates a textually pretrained large language model and adapts it into the audio-visual spoken dialogue domain by incorporating speech-text joint pretraining. Through extensive experiments, we validate the effectiveness of our model in facilitating a face-to-face conversation. Demo and data are available at https://multidialog.github.io and https://huggingface.co/datasets/IVLLab/MultiDialog, respectively.
URLs: https://multidialog.github.io, https://huggingface.co/datasets/IVLLab/MultiDialog,
Authors: Kepan Nan, Rui Xie, Penghao Zhou, Tiehan Fan, Zhenheng Yang, Zhijie Chen, Xiang Li, Jian Yang, Ying Tai
Abstract: Text-to-video (T2V) generation has recently garnered significant attention thanks to the large multi-modality model Sora. However, T2V generation still faces two important challenges: 1) Lacking a precise open sourced high-quality dataset. The previous popular video datasets, e.g. WebVid-10M and Panda-70M, are either with low quality or too large for most research institutions. Therefore, it is challenging but crucial to collect a precise high-quality text-video pairs for T2V generation. 2) Ignoring to fully utilize textual information. Recent T2V methods have focused on vision transformers, using a simple cross attention module for video generation, which falls short of thoroughly extracting semantic information from text prompt. To address these issues, we introduce OpenVid-1M, a precise high-quality dataset with expressive captions. This open-scenario dataset contains over 1 million text-video pairs, facilitating research on T2V generation. Furthermore, we curate 433K 1080p videos from OpenVid-1M to create OpenVidHD-0.4M, advancing high-definition video generation. Additionally, we propose a novel Multi-modal Video Diffusion Transformer (MVDiT) capable of mining both structure information from visual tokens and semantic information from text tokens. Extensive experiments and ablation studies verify the superiority of OpenVid-1M over previous datasets and the effectiveness of our MVDiT.
Authors: Sunny Gupta, Amit Sethi
Abstract: Federated Learning (FL) offers a privacy-preserving approach to train models on decentralized data. Its potential in healthcare is significant, but challenges arise due to cross-client variations in medical image data, exacerbated by limited annotations. This paper introduces Cross-Client Variations Adaptive Federated Learning (CCVA-FL) to address these issues. CCVA-FL aims to minimize cross-client variations by transforming images into a common feature space. It involves expert annotation of a subset of images from each client, followed by the selection of a client with the least data complexity as the target. Synthetic medical images are then generated using Scalable Diffusion Models with Transformers (DiT) based on the target client's annotated images. These synthetic images, capturing diversity and representing the original data, are shared with other clients. Each client then translates its local images into the target image space using image-to-image translation. The translated images are subsequently used in a federated learning setting to develop a server model. Our results demonstrate that CCVA-FL outperforms Vanilla Federated Averaging by effectively addressing data distribution differences across clients without compromising privacy.
Authors: Ming-Feng Li, Yueh-Feng Ku, Hong-Xuan Yen, Chi Liu, Yu-Lun Liu, Albert Y. C. Chen, Cheng-Hao Kuo, Min Sun
Abstract: Sparse RGBD scene completion is a challenging task especially when considering consistent textures and geometries throughout the entire scene. Different from existing solutions that rely on human-designed text prompts or predefined camera trajectories, we propose GenRC, an automated training-free pipeline to complete a room-scale 3D mesh with high-fidelity textures. To achieve this, we first project the sparse RGBD images to a highly incomplete 3D mesh. Instead of iteratively generating novel views to fill in the void, we utilized our proposed E-Diffusion to generate a view-consistent panoramic RGBD image which ensures global geometry and appearance consistency. Furthermore, we maintain the input-output scene stylistic consistency through textual inversion to replace human-designed text prompts. To bridge the domain gap among datasets, E-Diffusion leverages models trained on large-scale datasets to generate diverse appearances. GenRC outperforms state-of-the-art methods under most appearance and geometric metrics on ScanNet and ARKitScenes datasets, even though GenRC is not trained on these datasets nor using predefined camera trajectories. Project page: https://minfenli.github.io/GenRC
Authors: Tianxiao Zhang, Wenju Xu, Bo Luo, Guanghui Wang
Abstract: The Vision Transformer (ViT) leverages the Transformer's encoder to capture global information by dividing images into patches and achieves superior performance across various computer vision tasks. However, the self-attention mechanism of ViT captures the global context from the outset, overlooking the inherent relationships between neighboring pixels in images or videos. Transformers mainly focus on global information while ignoring the fine-grained local details. Consequently, ViT lacks inductive bias during image or video dataset training. In contrast, convolutional neural networks (CNNs), with their reliance on local filters, possess an inherent inductive bias, making them more efficient and quicker to converge than ViT with less data. In this paper, we present a lightweight Depth-Wise Convolution module as a shortcut in ViT models, bypassing entire Transformer blocks to ensure the models capture both local and global information with minimal overhead. Additionally, we introduce two architecture variants, allowing the Depth-Wise Convolution modules to be applied to multiple Transformer blocks for parameter savings, and incorporating independent parallel Depth-Wise Convolution modules with different kernels to enhance the acquisition of local information. The proposed approach significantly boosts the performance of ViT models on image classification, object detection and instance segmentation by a large margin, especially on small datasets, as evaluated on CIFAR-10, CIFAR-100, Tiny-ImageNet and ImageNet for image classification, and COCO for object detection and instance segmentation. The source code can be accessed at https://github.com/ZTX-100/Efficient_ViT_with_DW.
Authors: Haifeng Zhao, Lixiang Jiang, Leilei Ma, Dengdi Sun, Yanping Fu
Abstract: Medical images from different healthcare centers exhibit varied data distributions, posing significant challenges for adapting lung nodule detection due to the domain shift between training and application phases. Traditional unsupervised domain adaptive detection methods often struggle with this shift, leading to suboptimal outcomes. To overcome these challenges, we introduce a novel domain adaptive approach for lung nodule detection that leverages mean teacher self-training and contrastive learning. First, we propose a hierarchical contrastive learning strategy to refine nodule representations and enhance the distinction between nodules and background. Second, we introduce a nodule-level domain-invariant feature learning (NDL) module to capture domain-invariant features through adversarial learning across different domains. Additionally, we propose a new annotated dataset of X-ray images to aid in advancing lung nodule detection research. Extensive experiments conducted on multiple X-ray datasets demonstrate the efficacy of our approach in mitigating domain shift impacts.
Authors: Biao Wu, Yutong Xie, Zeyu Zhang, Minh Hieu Phan, Qi Chen, Ling Chen, Qi Wu
Abstract: Vision-and-language pretraining (VLP) in the medical field utilizes contrastive learning on image-text pairs to achieve effective transfer across tasks. Yet, current VLP approaches with the masked modelling strategy face two challenges when applied to the medical domain. First, current models struggle to accurately reconstruct key pathological features due to the scarcity of medical data. Second, most methods only adopt either paired image-text or image-only data, failing to exploit the combination of both paired and unpaired data. To this end, this paper proposes a XLIP (Masked modelling for medical Language-Image Pre-training) framework to enhance pathological learning and feature learning via unpaired data. First, we introduce the attention-masked image modelling (AttMIM) and entity-driven masked language modelling module (EntMLM), which learns to reconstruct pathological visual and textual tokens via multi-modal feature interaction, thus improving medical-enhanced features. The AttMIM module masks a portion of the image features that are highly responsive to textual features. This allows XLIP to improve the reconstruction of highly similar image data in medicine efficiency. Second, our XLIP capitalizes unpaired data to enhance multimodal learning by introducing disease-kind prompts. The experimental results show that XLIP achieves SOTA for zero-shot and fine-tuning classification performance on five datasets. Our code will be available at https://github.com/White65534/XLIP
Authors: Sudhir Yarram, Junsong Yuan
Abstract: Video extrapolation in space and time (VEST) enables viewers to forecast a 3D scene into the future and view it from novel viewpoints. Recent methods propose to learn an entangled representation, aiming to model layered scene geometry, motion forecasting and novel view synthesis together, while assuming simplified affine motion and homography-based warping at each scene layer, leading to inaccurate video extrapolation. Instead of entangled scene representation and rendering, our approach chooses to disentangle scene geometry from scene motion, via lifting the 2D scene to 3D point clouds, which enables high quality rendering of future videos from novel views. To model future 3D scene motion, we propose a disentangled two-stage approach that initially forecasts ego-motion and subsequently the residual motion of dynamic objects (e.g., cars, people). This approach ensures more precise motion predictions by reducing inaccuracies from entanglement of ego-motion with dynamic object motion, where better ego-motion forecasting could significantly enhance the visual outcomes. Extensive experimental analysis on two urban scene datasets demonstrate superior performance of our proposed method in comparison to strong baselines.
Authors: Chunlin Li, Hanrui Fan, Xiaorui Huang, Ruofan Liang, Sankeerth Durvasula, Nandita Vijaykumar
Abstract: We present a framework, DISORF, to enable online 3D reconstruction and visualization of scenes captured by resource-constrained mobile robots and edge devices. To address the limited computing capabilities of edge devices and potentially limited network availability, we design a framework that efficiently distributes computation between the edge device and the remote server. We leverage on-device SLAM systems to generate posed keyframes and transmit them to remote servers that can perform high-quality 3D reconstruction and visualization at runtime by leveraging recent advances in neural 3D methods. We identify a key challenge with online training where naive image sampling strategies can lead to significant degradation in rendering quality. We propose a novel shifted exponential frame sampling method that addresses this challenge for online training. We demonstrate the effectiveness of our framework in enabling high-quality real-time reconstruction and visualization of unknown scenes as they are captured and streamed from cameras in mobile robots and edge devices.
Authors: Joseph Shenouda, Yamin Zhou, Robert D. Nowak
Abstract: Motivated by the growing theoretical understanding of neural networks that employ the Rectified Linear Unit (ReLU) as their activation function, we revisit the use of ReLU activation functions for learning implicit neural representations (INRs). Inspired by second order B-spline wavelets, we incorporate a set of simple constraints to the ReLU neurons in each layer of a deep neural network (DNN) to remedy the spectral bias. This in turn enables its use for various INR tasks. Empirically, we demonstrate that, contrary to popular belief, one can learn state-of-the-art INRs based on a DNN composed of only ReLU neurons. Next, by leveraging recent theoretical works which characterize the kinds of functions ReLU neural networks learn, we provide a way to quantify the regularity of the learned function. This offers a principled approach to selecting the hyperparameters in INR architectures. We substantiate our claims through experiments in signal representation, super resolution, and computed tomography, demonstrating the versatility and effectiveness of our method. The code for all experiments can be found at https://github.com/joeshenouda/relu-inrs.
Authors: Hao Liang, Chengjie, Kun Li, Xin Tian
Abstract: Hyperspectral image (HSI) denoising is an essential procedure for HSI applications. Unfortunately, the existing Transformer-based methods mainly focus on non-local modeling, neglecting the importance of locality in image denoising. Moreover, deep learning methods employ complex spectral learning mechanisms, thus introducing large computation costs. To address these problems, we propose a hybrid spatial-spectral denoising network (HSSD), in which we design a novel hybrid dual-path network inspired by CNN and Transformer characteristics, leading to capturing both local and non-local spatial details while suppressing noise efficiently. Furthermore, to reduce computational complexity, we adopt a simple but effective decoupling strategy that disentangles the learning of space and spectral channels, where multilayer perception with few parameters is utilized to learn the global correlations among spectra. The synthetic and real experiments demonstrate that our proposed method outperforms state-of-the-art methods on spatial and spectral reconstruction. The code and details are available on https://github.com/HLImg/HSSD.
Authors: Jixiang Luo
Abstract: Since LIC has made rapid progress recently compared to traditional methods, this paper attempts to discuss the question about 'Where is the boundary of Learned Image Compression(LIC)?'. Thus this paper splits the above problem into two sub-problems:1)Where is the boundary of rate-distortion performance of PSNR? 2)How to further improve the compression gain and achieve the boundary? Therefore this paper analyzes the effectiveness of scaling parameters for encoder, decoder and context model, which are the three components of LIC. Then we conclude that scaling for LIC is to scale for context model and decoder within LIC. Extensive experiments demonstrate that overfitting can actually serve as an effective context. By optimizing the context, this paper further improves PSNR and achieves state-of-the-art performance, showing a performance gain of 14.39% with BD-RATE over VVC.
Authors: Xi Chen, Rahul Bhadani, Larry Head
Abstract: Current research on trajectory prediction primarily relies on data collected by onboard sensors of an ego vehicle. With the rapid advancement in connected technologies, such as vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication, valuable information from alternate views becomes accessible via wireless networks. The integration of information from alternative views has the potential to overcome the inherent limitations associated with a single viewpoint, such as occlusions and limited field of view. In this work, we introduce V2INet, a novel trajectory prediction framework designed to model multi-view data by extending existing single-view models. Unlike previous approaches where the multi-view data is manually fused or formulated as a separate training stage, our model supports end-to-end training, enhancing both flexibility and performance. Moreover, the predicted multimodal trajectories are calibrated by a post-hoc conformal prediction module to get valid and efficient confidence regions. We evaluated the entire framework using the real-world V2I dataset V2X-Seq. Our results demonstrate superior performance in terms of Final Displacement Error (FDE) and Miss Rate (MR) using a single GPU. The code is publicly available at: \url{https://github.com/xichennn/V2I_trajectory_prediction}.