Authors: Anton Jeran Ratnarajah, Sahani Goonetilleke, Dumindu Tissera, Kapilan Balagopalan, Ranga Rodrigo
Law enforcement officials heavily depend on Forensic Video Analytic (FVA) Software in their evidence extraction process. However present-day FVA software are complex, time consuming, equipment dependent and expensive. Developing countries struggle to gain access to this gateway to a secure haven. The term forensic pertains the application of scientific methods to the investigation of crime through post-processing, whereas surveillance is the close monitoring of real-time feeds.
The principle objective of this Final Year Project was to develop an efficient and effective FVA Software, addressing the shortcomings through a stringent and systematic review of scholarly research papers, online databases and legal documentation. The scope spans multiple object detection, multiple object tracking, anomaly detection, activity recognition, tampering detection, general and specific image enhancement and video synopsis.
Methods employed include many machine learning techniques, GPU acceleration and efficient, integrated architecture development both for real-time and postprocessing. For this CNN, GMM, multithreading and OpenCV C++ coding were used. The implications of the proposed methodology would rapidly speed up the FVA process especially through the novel video synopsis research arena. This project has resulted in three research outcomes Moving Object Based Collision Free Video Synopsis, Forensic and Surveillance Analytic Tool Architecture and Tampering Detection Inter-Frame Forgery.
The results include forensic and surveillance panel outcomes with emphasis on video synopsis and Sri Lankan context. Principal conclusions include the optimization and efficient algorithm integration to overcome limitations in processing power, memory and compromise between real-time performance and accuracy.
Authors: Manna Dai, Yang Jiang, Feng Yang, Joyjit Chattoraj, Yingzhi Xia, Xinxing Xu, Weijiang Zhao, My Ha Dao, Yong Liu
Metasurfaces have widespread applications in fifth-generation (5G) microwave communication. Among the metasurface family, free-form metasurfaces excel in achieving intricate spectral responses compared to regular-shape counterparts. However, conventional numerical methods for free-form metasurfaces are time-consuming and demand specialized expertise. Alternatively, recent studies demonstrate that deep learning has great potential to accelerate and refine metasurface designs. Here, we present XGAN, an extended generative adversarial network (GAN) with a surrogate for high-quality free-form metasurface designs. The proposed surrogate provides a physical constraint to XGAN so that XGAN can accurately generate metasurfaces monolithically from input spectral responses. In comparative experiments involving 20000 free-form metasurface designs, XGAN achieves 0.9734 average accuracy and is 500 times faster than the conventional methodology. This method facilitates the metasurface library building for specific spectral responses and can be extended to various inverse design problems, including optical metamaterials, nanophotonic devices, and drug discovery.
Authors: Vahid Mohammadi Safarzadeh
Vessel structure is one of the most important parts of the retina which physicians can detect many diseases by analysing its features. Localization of blood vessels in retina images is an important process in medical image analysis. This process is also more challenging with the presence of bright and dark lesions. In this thesis, two automated vessel localization methods to handle both healthy and unhealthy (pathological) retina images are analyzed. Each method consists of two major steps and the second step is the same in the two methods. In the first step, an algorithm is used to decrease the effect of bright lesions. In Method 1, this algorithm is based on K- Means segmentation, and in Method 2, it is based on a regularization procedure. In the second step of both methods, a multi-scale line operator is used to localize the line-shaped vascular structures and ignore the dark lesions which are generally assumed to have irregular patterns. After the introduction of the methods, a detailed quantitative and qualitative comparison of the methods with one another as well as the state-of-the-art solutions in the literature based on the segmentation results on the images of the two publicly available datasets, DRIVE and STARE, is reported. The results demonstrate that the methods are highly comparable with other solutions.
Authors: Yuntao Wei, Yuzhe Zhang, Shuyang Zhang, Hong Zhang
Multimodal depression detection is an important research topic that aims to predict human mental states using multimodal data. Previous methods treat different modalities equally and fuse each modality by na\"ive mathematical operations without measuring the relative importance between them, which cannot obtain well-performed multimodal representations for downstream depression tasks. In order to tackle the aforementioned concern, we present a Cross-modal Attention Network with Adaptive Multi-modal Recurrent Fusion (CANAMRF) for multimodal depression detection. CANAMRF is constructed by a multimodal feature extractor, an Adaptive Multimodal Recurrent Fusion module, and a Hybrid Attention Module. Through experimentation on two benchmark datasets, CANAMRF demonstrates state-of-the-art performance, underscoring the effectiveness of our proposed approach.
Authors: Siyuan Yan, Chi Liu, Zhen Yu, Lie Ju, Dwarikanath Mahapatra, Brigid Betz-Stablein, Victoria Mar, Monika Janda, Peter Soyer, Zongyuan Ge
Deep learning models for medical image analysis easily suffer from distribution shifts caused by dataset artifacts bias, camera variations, differences in the imaging station, etc., leading to unreliable diagnoses in real-world clinical settings. Domain generalization (DG) methods, which aim to train models on multiple domains to perform well on unseen domains, offer a promising direction to solve the problem. However, existing DG methods assume domain labels of each image are available and accurate, which is typically feasible for only a limited number of medical datasets. To address these challenges, we propose a novel DG framework for medical image classification without relying on domain labels, called Prompt-driven Latent Domain Generalization (PLDG). PLDG consists of unsupervised domain discovery and prompt learning. This framework first discovers pseudo domain labels by clustering the bias-associated style features, then leverages collaborative domain prompts to guide a Vision Transformer to learn knowledge from discovered diverse domains. To facilitate cross-domain knowledge learning between different prompts, we introduce a domain prompt generator that enables knowledge sharing between domain prompts and a shared prompt. A domain mixup strategy is additionally employed for more flexible decision margins and mitigates the risk of incorrect domain assignments. Extensive experiments on three medical image classification tasks and one debiasing task demonstrate that our method can achieve comparable or even superior performance than conventional DG algorithms without relying on domain labels. Our code will be publicly available upon the paper is accepted.
Authors: Sudipan Saha, Tushar Verma, Dario Augusto Borges Oliveira
With the global population on the rise, our cities have been expanding to accommodate the growing number of people. The expansion of cities generally leads to the engulfment of peripheral areas. However, such expansion of urban areas is likely to cause increment in areas with increased land surface temperature (LST). By considering each summer as a data point, we form LST multi-year time-series and cluster it to obtain spatio-temporal pattern. We observe several interesting phenomena from these patterns, e.g., some clusters show reasonable similarity to the built-up area, whereas the locations with high temporal variation are seen more in the peripheral areas. Furthermore, the LST center of mass shifts over the years for cities with development activities tilted towards a direction. We conduct the above-mentioned studies for three different cities in three different continents.
Authors: Niloufar Alipour Talemi, Hossein Kashiani, Nasser M. Nasrabadi
Although face recognition (FR) has achieved great success in recent years, it is still challenging to accurately recognize faces in low-quality images due to the obscured facial details. Nevertheless, it is often feasible to make predictions about specific soft biometric (SB) attributes, such as gender, and baldness even in dealing with low-quality images. In this paper, we propose a novel multi-branch neural network that leverages SB attribute information to boost the performance of FR. To this end, we propose a cross-attribute-guided transformer fusion (CATF) module that effectively captures the long-range dependencies and relationships between FR and SB feature representations. The synergy created by the reciprocal flow of information in the dual cross-attention operations of the proposed CATF module enhances the performance of FR. Furthermore, we introduce a novel self-attention distillation framework that effectively highlights crucial facial regions, such as landmarks by aligning low-quality images with those of their high-quality counterparts in the feature space. The proposed self-attention distillation regularizes our network to learn a unified quality-invariant feature representation in unconstrained environments. We conduct extensive experiments on various FR benchmarks varying in quality. Experimental results demonstrate the superiority of our FR method compared to state-of-the-art FR studies.
Authors: Kebin Wu, Wenbin Li, Xiaofei Xiao
Traffic accident analysis is pivotal for enhancing public safety and developing road regulations. Traditional approaches, although widely used, are often constrained by manual analysis processes, subjective decisions, uni-modal outputs, as well as privacy issues related to sensitive data. This paper introduces the idea of AccidentGPT, a foundation model of traffic accident analysis, which incorporates multi-modal input data to automatically reconstruct the accident process video with dynamics details, and furthermore provide multi-task analysis with multi-modal outputs. The design of the AccidentGPT is empowered with a multi-modality prompt with feedback for task-oriented adaptability, a hybrid training schema to leverage labelled and unlabelled data, and a edge-cloud split configuration for data privacy. To fully realize the functionalities of this model, we proposes several research opportunities. This paper serves as the stepping stone to fill the gaps in traditional approaches of traffic accident analysis and attract the research community attention for automatic, objective, and privacy-preserving traffic accident analysis.
Authors: Qihua Chen, Xuejin Chen, Chenxuan Wang, Yixiong Liu, Zhiwei Xiong, Feng Wu
The current neuron reconstruction pipeline for electron microscopy (EM) data usually includes automatic image segmentation followed by extensive human expert proofreading. In this work, we aim to reduce human workload by predicting connectivity between over-segmented neuron pieces, taking both microscopy image and 3D morphology features into account, similar to human proofreading workflow. To this end, we first construct a dataset, named FlyTracing, that contains millions of pairwise connections of segments expanding the whole fly brain, which is three orders of magnitude larger than existing datasets for neuron segment connection. To learn sophisticated biological imaging features from the connectivity annotations, we propose a novel connectivity-aware contrastive learning method to generate dense volumetric EM image embedding. The learned embeddings can be easily incorporated with any point or voxel-based morphological representations for automatic neuron tracing. Extensive comparisons of different combination schemes of image and morphological representation in identifying split errors across the whole fly brain demonstrate the superiority of the proposed approach, especially for the locations that contain severe imaging artifacts, such as section missing and misalignment. The dataset and code are available at https://github.com/Levishery/Flywire-Neuron-Tracing.
Authors: Xin Ma, Yaohui Wang, Gengyun Jia, Xinyuan Chen, Ziwei Liu, Yuan-Fang Li, Cunjian Chen, Yu Qiao
We propose a novel Latent Diffusion Transformer, namely Latte, for video generation. Latte first extracts spatio-temporal tokens from input videos and then adopts a series of Transformer blocks to model video distribution in the latent space. In order to model a substantial number of tokens extracted from videos, four efficient variants are introduced from the perspective of decomposing the spatial and temporal dimensions of input videos. To improve the quality of generated videos, we determine the best practices of Latte through rigorous experimental analysis, including video clip patch embedding, model variants, timestep-class information injection, temporal positional embedding, and learning strategies. Our comprehensive evaluation demonstrates that Latte achieves state-of-the-art performance across four standard video generation datasets, i.e., FaceForensics, SkyTimelapse, UCF101, and Taichi-HD. In addition, we extend Latte to text-to-video generation (T2V) task, where Latte achieves comparable results compared to recent T2V models. We strongly believe that Latte provides valuable insights for future research on incorporating Transformers into diffusion models for video generation.
Authors: Ho Hin Lee, Adam M. Saunders, Michael E. Kim, Samuel W. Remedios, Yucheng Tang, Qi Yang, Xin Yu, Shunxing Bao, Chloe Cho, Louise A. Mawn, Tonia S. Rex, Kevin L. Schey, Blake E. Dewey, Jeffrey M. Spraggins, Jerry L. Prince, Yuankai Huo, Bennett A. Landman
Eye morphology varies significantly across the population, especially for the orbit and optic nerve. These variations limit the feasibility and robustness of generalizing population-wise features of eye organs to an unbiased spatial reference. To tackle these limitations, we propose a process for creating high-resolution unbiased eye atlases. First, to restore spatial details from scans with a low through-plane resolution compared to a high in-plane resolution, we apply a deep learning-based super-resolution algorithm. Then, we generate an initial unbiased reference with an iterative metric-based registration using a small portion of subject scans. We register the remaining scans to this template and refine the template using an unsupervised deep probabilistic approach that generates a more expansive deformation field to enhance the organ boundary alignment. We demonstrate this framework using magnetic resonance images across four different MRI tissue contrasts, generating four atlases in separate spatial alignments. For each tissue contrast, we find a significant improvement in the average Dice score across four labeled regions compared to a standard registration framework consisting of rigid, affine, and deformable transformations. These results highlight the effective alignment of eye organs and boundaries using our proposed process. By combining super-resolution preprocessing and deep probabilistic models, we address the challenge of generating an eye atlas to serve as a standardized reference across a largely variable population.
Authors: Geoffery Agorku, Sarah Hernandez PhD, Maria Falquez, Subhadipto Poddar PhD, Kwadwo Amankwah-Nkyi
Inland waterways are critical for freight movement, but limited means exist for monitoring their performance and usage by freight-carrying vessels, e.g., barges. While methods to track vessels, e.g., tug and tow boats, are publicly available through Automatic Identification Systems (AIS), ways to track freight tonnages and commodity flows carried on barges along these critical marine highways are non-existent, especially in real-time settings. This paper develops a method to detect barge traffic on inland waterways using existing traffic cameras with opportune viewing angles. Deep learning models, specifically, You Only Look Once (YOLO), Single Shot MultiBox Detector (SSD), and EfficientDet are employed. The model detects the presence of vessels and/or barges from video and performs a classification (no vessel or barge, vessel without barge, vessel with barge, and barge). A dataset of 331 annotated images was collected from five existing traffic cameras along the Mississippi and Ohio Rivers for model development. YOLOv8 achieves an F1-score of 96%, outperforming YOLOv5, SSD, and EfficientDet models with 86%, 79%, and 77% respectively. Sensitivity analysis was carried out regarding weather conditions (fog and rain) and location (Mississippi and Ohio rivers). A background subtraction technique was used to normalize video images across the various locations for the location sensitivity analysis. This model can be used to detect the presence of barges along river segments, which can be used for anonymous bulk commodity tracking and monitoring. Such data is valuable for long-range transportation planning efforts carried out by public transportation agencies, in addition to operational and maintenance planning conducted by federal agencies such as the US Army Corp of Engineers.
Authors: Paul Brimoh, Chollette C. Olisah
A genuine signer's signature is naturally unstable even at short time-intervals whereas, expert forgers always try to perfectly mimic a genuine signer's signature. This presents a challenge which puts a genuine signer at risk of being denied access, while a forge signer is granted access. The implication is a high false acceptance rate (FAR) which is the percentage of forge signature classified as belonging to a genuine class. Existing work have only scratched the surface of signature verification because the misclassification error remains high. In this paper, a consensus-threshold distance-based classifier criterion is proposed for offline writer-dependent signature verification. Using features extracted from SigNet and SigNet-F deep convolutional neural network models, the proposed classifier minimizes FAR. This is demonstrated via experiments on four datasets: GPDS-300, MCYT, CEDAR and Brazilian PUC-PR datasets. On GPDS-300, the consensus threshold classifier improves the state-of-the-art performance by achieving a 1.27% FAR compared to 8.73% and 17.31% recorded in literature. This performance is consistent across other datasets and guarantees that the risk of imposters gaining access to sensitive documents or transactions is minimal.
Authors: Xin He, Longhui Wei, Lingxi Xie, Qi Tian
Multimodal Large Language Models (MLLMs) are experiencing rapid growth, yielding a plethora of noteworthy contributions in recent months. The prevailing trend involves adopting data-driven methodologies, wherein diverse instruction-following datasets are collected. However, a prevailing challenge persists in these approaches, specifically in relation to the limited visual perception ability, as CLIP-like encoders employed for extracting visual information from inputs. Though these encoders are pre-trained on billions of image-text pairs, they still grapple with the information loss dilemma, given that textual captions only partially capture the contents depicted in images. To address this limitation, this paper proposes to improve the visual perception ability of MLLMs through a mixture-of-experts knowledge enhancement mechanism. Specifically, we introduce a novel method that incorporates multi-task encoders and visual tools into the existing MLLMs training and inference pipeline, aiming to provide a more comprehensive and accurate summarization of visual inputs. Extensive experiments have evaluated its effectiveness of advancing MLLMs, showcasing improved visual perception achieved through the integration of visual experts.
Authors: Shanthika Naik, Kunwar Singh, Astitva Srivastava, Dhawal Sirikonda, Amit Raj, Varun Jampani, Avinash Sharma
We propose a novel self-supervised framework for retargeting non-parameterized 3D garments onto 3D human avatars of arbitrary shapes and poses, enabling 3D virtual try-on (VTON). Existing self-supervised 3D retargeting methods only support parametric and canonical garments, which can only be draped over parametric body, e.g. SMPL. To facilitate the non-parametric garments and body, we propose a novel method that introduces Isomap Embedding based correspondences matching between the garment and the human body to get a coarse alignment between the two meshes. We perform neural refinement of the coarse alignment in a self-supervised setting. Further, we leverage a Laplacian detail integration method for preserving the inherent details of the input garment. For evaluating our 3D non-parametric garment retargeting framework, we propose a dataset of 255 real-world garments with realistic noise and topological deformations. The dataset contains $44$ unique garments worn by 15 different subjects in 5 distinctive poses, captured using a multi-view RGBD capture setup. We show superior retargeting quality on non-parametric garments and human avatars over existing state-of-the-art methods, acting as the first-ever baseline on the proposed dataset for non-parametric 3D garment retargeting.
Authors: Yang Sui, Zhuohang Li, Ding Ding, Xiang Pan, Xiaozhong Xu, Shan Liu, Zhenzhong Chen
Adversarial attacks can readily disrupt the image classification system, revealing the vulnerability of DNN-based recognition tasks. While existing adversarial perturbations are primarily applied to uncompressed images or compressed images by the traditional image compression method, i.e., JPEG, limited studies have investigated the robustness of models for image classification in the context of DNN-based image compression. With the rapid evolution of advanced image compression, DNN-based learned image compression has emerged as the promising approach for transmitting images in many security-critical applications, such as cloud-based face recognition and autonomous driving, due to its superior performance over traditional compression. Therefore, there is a pressing need to fully investigate the robustness of a classification system post-processed by learned image compression. To bridge this research gap, we explore the adversarial attack on a new pipeline that targets image classification models that utilize learned image compressors as pre-processing modules. Furthermore, to enhance the transferability of perturbations across various quality levels and architectures of learned image compression models, we introduce a saliency score-based sampling method to enable the fast generation of transferable perturbation. Extensive experiments with popular attack methods demonstrate the enhanced transferability of our proposed method when attacking images that have been post-processed with different learned image compression models.
Authors: Xuran Hu, Ziqiang Xu, Zhihan Chen, Zhengpeng Feng, Mingzhe Zhu, LJubisa Stankovic
Speckle noise poses a significant challenge in maintaining the quality of synthetic aperture radar (SAR) images, so SAR despeckling techniques have drawn increasing attention. Despite the tremendous advancements of deep learning in fixed-scale SAR image despeckling, these methods still struggle to deal with large-scale SAR images. To address this problem, this paper introduces a novel despeckling approach termed Region Denoising Diffusion Probabilistic Model (R-DDPM) based on generative models. R-DDPM enables versatile despeckling of SAR images across various scales, accomplished within a single training session. Moreover, The artifacts in the fused SAR images can be avoided effectively with the utilization of region-guided inverse sampling. Experiments of our proposed R-DDPM on Sentinel-1 data demonstrates superior performance to existing methods.
Authors: Junhuan Yang, Hanchen Wang, Yi Sheng, Youzuo Lin, Lei Yang
Full-waveform inversion (FWI) plays a vital role in geoscience to explore the subsurface. It utilizes the seismic wave to image the subsurface velocity map. As the machine learning (ML) technique evolves, the data-driven approaches using ML for FWI tasks have emerged, offering enhanced accuracy and reduced computational cost compared to traditional physics-based methods. However, a common challenge in geoscience, the unprivileged data, severely limits ML effectiveness. The issue becomes even worse during model pruning, a step essential in geoscience due to environmental complexities. To tackle this, we introduce the EdGeo toolkit, which employs a diffusion-based model guided by physics principles to generate high-fidelity velocity maps. The toolkit uses the acoustic wave equation to generate corresponding seismic waveform data, facilitating the fine-tuning of pruned ML models. Our results demonstrate significant improvements in SSIM scores and reduction in both MAE and MSE across various pruning ratios. Notably, the ML model fine-tuned using data generated by EdGeo yields superior quality of velocity maps, especially in representing unprivileged features, outperforming other existing methods.
Authors: Taymaz Akan, Sait Alp, Mohammad A. N Bhuiyanb
Alzheimer's is a brain disease that gets worse over time and affects memory, thinking, and behavior. Alzheimer's disease (AD) can be treated and managed if it is diagnosed early, which can slow the progression of symptoms and improve quality of life. In this study, we suggested using the Visual Transformer (ViT) and bi-LSTM to process MRI images for diagnosing Alzheimer's disease. We used ViT to extract features from the MRI and then map them to a feature sequence. Then, we used Bi-LSTM sequence modeling to keep the interdependencies between related features. In addition, we evaluated the performance of the proposed model for the binary classification of AD patients using data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Finally, we evaluated our method against other deep learning models in the literature. The proposed method performs well in terms of accuracy, precision, F-score, and recall for the diagnosis of AD.
Authors: Liangtao Shi, Bineng Zhong, Qihua Liang, Ning Li, Shengping Zhang, Xianxian Li
How to effectively exploit spatio-temporal information is crucial to capture target appearance changes in visual tracking. However, most deep learning-based trackers mainly focus on designing a complicated appearance model or template updating strategy, while lacking the exploitation of context between consecutive frames and thus entailing the \textit{when-and-how-to-update} dilemma. To address these issues, we propose a novel explicit visual prompts framework for visual tracking, dubbed \textbf{EVPTrack}. Specifically, we utilize spatio-temporal tokens to propagate information between consecutive frames without focusing on updating templates. As a result, we cannot only alleviate the challenge of \textit{when-to-update}, but also avoid the hyper-parameters associated with updating strategies. Then, we utilize the spatio-temporal tokens to generate explicit visual prompts that facilitate inference in the current frame. The prompts are fed into a transformer encoder together with the image tokens without additional processing. Consequently, the efficiency of our model is improved by avoiding \textit{how-to-update}. In addition, we consider multi-scale information as explicit visual prompts, providing multiscale template features to enhance the EVPTrack's ability to handle target scale changes. Extensive experimental results on six benchmarks (i.e., LaSOT, LaSOT\rm $_{ext}$, GOT-10k, UAV123, TrackingNet, and TNL2K.) validate that our EVPTrack can achieve competitive performance at a real-time speed by effectively exploiting both spatio-temporal and multi-scale information. Code and models are available at https://github.com/GXNU-ZhongLab/EVPTrack.
Authors: Yuanpeng Tu, Boshen Zhang, Liang Liu, Yuxi Li, Chenhai Xu, Jiangning Zhang, Yabiao Wang, Chengjie Wang, Cai Rong Zhao
Industrial anomaly detection is generally addressed as an unsupervised task that aims at locating defects with only normal training samples. Recently, numerous 2D anomaly detection methods have been proposed and have achieved promising results, however, using only the 2D RGB data as input is not sufficient to identify imperceptible geometric surface anomalies. Hence, in this work, we focus on multi-modal anomaly detection. Specifically, we investigate early multi-modal approaches that attempted to utilize models pre-trained on large-scale visual datasets, i.e., ImageNet, to construct feature databases. And we empirically find that directly using these pre-trained models is not optimal, it can either fail to detect subtle defects or mistake abnormal features as normal ones. This may be attributed to the domain gap between target industrial data and source data.Towards this problem, we propose a Local-to-global Self-supervised Feature Adaptation (LSFA) method to finetune the adaptors and learn task-oriented representation toward anomaly detection.Both intra-modal adaptation and cross-modal alignment are optimized from a local-to-global perspective in LSFA to ensure the representation quality and consistency in the inference stage.Extensive experiments demonstrate that our method not only brings a significant performance boost to feature embedding based approaches, but also outperforms previous State-of-The-Art (SoTA) methods prominently on both MVTec-3D AD and Eyecandies datasets, e.g., LSFA achieves 97.1% I-AUROC on MVTec-3D, surpass previous SoTA by +3.4%.
Authors: Yixin Chen, Shuai Zhang, Boran Han, Tong He, Bo Li
In this work, we introduce Context-Aware MultiModal Learner (CaMML), for tuning large multimodal models (LMMs). CaMML, a lightweight module, is crafted to seamlessly integrate multimodal contextual samples into large models, thereby empowering the model to derive knowledge from analogous, domain-specific, up-to-date information and make grounded inferences. Importantly, CaMML is highly scalable and can efficiently handle lengthy multimodal context examples owing to its hierarchical design. Based on CaMML, we have developed two multimodal models, CaMML-7B and CaMML-13B, that have shown exceptional performance across an array of benchmark datasets for multimodal tasks. Remarkably, CaMML-13B achieves the state-of-the-art performance on over ten widely recognized multimodal benchmark datasets, surpassing LLaVA-1.5 (13B) with a noticeable margin, without integration of any external resources. Moreover, we have conducted extensive ablative studies to inspect the inner workings of CaMML and performed qualitative analyses to showcase its effectiveness in handling real-world challenging cases.
Authors: Gabriele Valvano, Antonino Agostino, Giovanni De Magistris, Antonino Graziano, Giacomo Veneri
Training supervised deep neural networks that perform defect detection and segmentation requires large-scale fully-annotated datasets, which can be hard or even impossible to obtain in industrial environments. Generative AI offers opportunities to enlarge small industrial datasets artificially, thus enabling the usage of state-of-the-art supervised approaches in the industry. Unfortunately, also good generative models need a lot of data to train, while industrial datasets are often tiny. Here, we propose a new approach for reusing general-purpose pre-trained generative models on industrial data, ultimately allowing the generation of self-labelled defective images. First, we let the model learn the new concept, entailing the novel data distribution. Then, we force it to learn to condition the generative process, producing industrial images that satisfy well-defined topological characteristics and show defects with a given geometry and location. To highlight the advantage of our approach, we use the synthetic dataset to optimise a crack segmentor for a real industrial use case. When the available data is small, we observe considerable performance increase under several metrics, showing the method's potential in production environments.
Authors: Bo Zhang, Yuqi Han, Jinli Suo, Qionghai Dai
Event cameras or dynamic vision sensors (DVS) record asynchronous response to brightness changes instead of conventional intensity frames, and feature ultra-high sensitivity at low bandwidth. The new mechanism demonstrates great advantages in challenging scenarios with fast motion and large dynamic range. However, the recorded events might be highly sparse due to either limited hardware bandwidth or extreme photon starvation in harsh environments. To unlock the full potential of event cameras, we propose an inventive event sequence completion approach conforming to the unique characteristics of event data in both the processing stage and the output form. Specifically, we treat event streams as 3D event clouds in the spatiotemporal domain, develop a diffusion-based generative model to generate dense clouds in a coarse-to-fine manner, and recover exact timestamps to maintain the temporal resolution of raw data successfully. To validate the effectiveness of our method comprehensively, we perform extensive experiments on three widely used public datasets with different spatial resolutions, and additionally collect a novel event dataset covering diverse scenarios with highly dynamic motions and under harsh illumination. Besides generating high-quality dense events, our method can benefit downstream applications such as object classification and intensity frame reconstruction.
Authors: Sahan Dissanayaka, Oshan Mudanayaka, Thilina Halloluwa, Chameera De Silva
Image processing holds immense potential for societal benefit, yet its full potential is often accessible only to tech-savvy experts. Bridging this knowledge gap and providing accessible tools for users of all backgrounds remains an unexplored frontier. This paper introduces "ImageLab," a novel tool designed to democratize image processing, catering to both novices and experts by prioritizing interactive learning over theoretical complexity. ImageLab not only serves as a valuable educational resource but also offers a practical testing environment for seasoned practitioners. Through a comprehensive evaluation of ImageLab's features, we demonstrate its effectiveness through a user study done for a focused group of school children and university students which enables us to get positive feedback on the tool. Our work represents a significant stride toward enhancing image processing education and practice, making it more inclusive and approachable for all.
Authors: Vibhu Dalal
Variational Autoencoders (VAEs) are powerful generative models, however their generated samples are known to suffer from a characteristic blurriness, as compared to the outputs of alternative generating techniques. Extensive research efforts have been made to tackle this problem, and several works have focused on modifying the reconstruction term of the evidence lower bound (ELBO). In particular, many have experimented with augmenting the reconstruction loss with losses in the frequency domain. Such loss functions usually employ the Fourier transform to explicitly penalise the lack of higher frequency components in the generated samples, which are responsible for sharp visual features. In this paper, we explore the aspects of previous such approaches which aren't well understood, and we propose an augmentation to the reconstruction term in response to them. Our reasoning leads us to use the short-time Fourier transform and to emphasise on local phase coherence between the input and output samples. We illustrate the potential of our proposed loss on the MNIST dataset by providing both qualitative and quantitative results.
Authors: Rui She, Sijie Wang, Qiyu Kang, Kai Zhao, Yang Song, Wee Peng Tay, Tianyu Geng, Xingchao Jian
Point cloud registration is a crucial technique in 3D computer vision with a wide range of applications. However, this task can be challenging, particularly in large fields of view with dynamic objects, environmental noise, or other perturbations. To address this challenge, we propose a model called PosDiffNet. Our approach performs hierarchical registration based on window-level, patch-level, and point-level correspondence. We leverage a graph neural partial differential equation (PDE) based on Beltrami flow to obtain high-dimensional features and position embeddings for point clouds. We incorporate position embeddings into a Transformer module based on a neural ordinary differential equation (ODE) to efficiently represent patches within points. We employ the multi-level correspondence derived from the high feature similarity scores to facilitate alignment between point clouds. Subsequently, we use registration methods such as SVD-based algorithms to predict the transformation using corresponding point pairs. We evaluate PosDiffNet on several 3D point cloud datasets, verifying that it achieves state-of-the-art (SOTA) performance for point cloud registration in large fields of view with perturbations. The implementation code of experiments is available at https://github.com/AI-IT-AVs/PosDiffNet.
Authors: Chujie Zhao, Tianren Zhang, Feng Chen
Domain generalization (DG) aims to improve the generalization ability of the model trained on several known training domains over unseen test domains. Previous work has shown that self-supervised contrastive pre-training improves the robustness of the model on downstream tasks. However, in this paper, we find that self-supervised models do not exhibit better generalization performance than supervised models pre-trained on the same dataset in the DG setting. We argue that this is owing to the fact that the richer intra-class discriminative features extracted by self-supervised contrastive learning, which we term silent features, are suppressed during supervised fine-tuning. These silent features are likely to contain features that are more generalizable on the test domain. In this work, we model and analyze this feature suppression phenomenon and theoretically prove that preserving silent features can achieve lower expected test domain risk under certain conditions. In light of this, we propose a simple yet effective method termed STEP (Silent Feature Preservation) to improve the generalization performance of the self-supervised contrastive learning pre-trained model by alleviating the suppression of silent features during the supervised fine-tuning process. Experimental results show that STEP exhibits state-of-the-art performance on standard DG benchmarks with significant distribution shifts.
Authors: Tran Cao Minh, Nguyen Kim Quoc, Phan Cong Vinh, Dang Nhu Phu, Vuong Xuan Chi, Ha Minh Tan
In the field of medical imaging, breast ultrasound has emerged as a crucial diagnostic tool for early detection of breast cancer. However, the accuracy of diagnosing the location of the affected area and the extent of the disease depends on the experience of the physician. In this paper, we propose a novel model called UGGNet, combining the power of the U-Net and VGG architectures to enhance the performance of breast ultrasound image analysis. The U-Net component of the model helps accurately segment the lesions, while the VGG component utilizes deep convolutional layers to extract features. The fusion of these two architectures in UGGNet aims to optimize both segmentation and feature representation, providing a comprehensive solution for accurate diagnosis in breast ultrasound images. Experimental results have demonstrated that the UGGNet model achieves a notable accuracy of 78.2% on the "Breast Ultrasound Images Dataset."
Authors: Qian Li, Lixin Su, Jiashu Zhao, Long Xia, Hengyi Cai, Suqi Cheng, Hengzhu Tang, Junfeng Wang, Dawei Yin
Text-video retrieval is a challenging task that aims to identify relevant videos given textual queries. Compared to conventional textual retrieval, the main obstacle for text-video retrieval is the semantic gap between the textual nature of queries and the visual richness of video content. Previous works primarily focus on aligning the query and the video by finely aggregating word-frame matching signals. Inspired by the human cognitive process of modularly judging the relevance between text and video, the judgment needs high-order matching signal due to the consecutive and complex nature of video contents. In this paper, we propose chunk-level text-video matching, where the query chunks are extracted to describe a specific retrieval unit, and the video chunks are segmented into distinct clips from videos. We formulate the chunk-level matching as n-ary correlations modeling between words of the query and frames of the video and introduce a multi-modal hypergraph for n-ary correlation modeling. By representing textual units and video frames as nodes and using hyperedges to depict their relationships, a multi-modal hypergraph is constructed. In this way, the query and the video can be aligned in a high-order semantic space. In addition, to enhance the model's generalization ability, the extracted features are fed into a variational inference component for computation, obtaining the variational representation under the Gaussian distribution. The incorporation of hypergraphs and variational inference allows our model to capture complex, n-ary interactions among textual and visual contents. Experimental results demonstrate that our proposed method achieves state-of-the-art performance on the text-video retrieval task.
Authors: Jiaqing Zhang, Jie Lei, Weiying Xie, Geng Yang, Daixun Li, Yunsong Li, Karim Seghouane
In multimodal land cover classification (MLCC), a common challenge is the redundancy in data distribution, where irrelevant information from multiple modalities can hinder the effective integration of their unique features. To tackle this, we introduce the Multimodal Informative Vit (MIVit), a system with an innovative information aggregate-distributing mechanism. This approach redefines redundancy levels and integrates performance-aware elements into the fused representation, facilitating the learning of semantics in both forward and backward directions. MIVit stands out by significantly reducing redundancy in the empirical distribution of each modality's separate and fused features. It employs oriented attention fusion (OAF) for extracting shallow local features across modalities in horizontal and vertical dimensions, and a Transformer feature extractor for extracting deep global features through long-range attention. We also propose an information aggregation constraint (IAC) based on mutual information, designed to remove redundant information and preserve complementary information within embedded features. Additionally, the information distribution flow (IDF) in MIVit enhances performance-awareness by distributing global classification information across different modalities' feature maps. This architecture also addresses missing modality challenges with lightweight independent modality classifiers, reducing the computational load typically associated with Transformers. Our results show that MIVit's bidirectional aggregate-distributing mechanism between modalities is highly effective, achieving an average overall accuracy of 95.56% across three multimodal datasets. This performance surpasses current state-of-the-art methods in MLCC. The code for MIVit is accessible at https://github.com/icey-zhang/MIViT.
Authors: Jiaqing Zhang, Jie Lei, Weiying Xie, Kai Jiang, Mingxiang Cao, Yunsong Li
Accurate cloud recognition and warning are crucial for various applications, including in-flight support, weather forecasting, and climate research. However, recent deep learning algorithms have predominantly focused on detecting cloud regions in satellite imagery, with insufficient attention to the specificity required for accurate cloud recognition. This limitation inspired us to develop the novel FY-4A-Himawari-8 (FYH) dataset, which includes nine distinct cloud categories and uses precise domain adaptation methods to align 70,419 image-label pairs in terms of projection, temporal resolution, and spatial resolution, thereby facilitating the training of supervised deep learning networks. Given the complexity and diversity of cloud formations, we have thoroughly analyzed the challenges inherent to cloud recognition tasks, examining the intricate characteristics and distribution of the data. To effectively address these challenges, we designed a Distribution-aware Interactive-Attention Network (DIAnet), which preserves pixel-level details through a high-resolution branch and a parallel multi-resolution cross-branch. We also integrated a distribution-aware loss (DAL) to mitigate the imbalance across cloud categories. An Interactive Attention Module (IAM) further enhances the robustness of feature extraction combined with spatial and channel information. Empirical evaluations on the FYH dataset demonstrate that our method outperforms other cloud recognition networks, achieving superior performance in terms of mean Intersection over Union (mIoU). The code for implementing DIAnet is available at https://github.com/icey-zhang/DIAnet.
Authors: Nianwen Si, Hao Zhang, Weiqiang Zhang
Large language models are known for encoding a vast amount of factual knowledge, but they often becomes outdated due to the ever-changing nature of external information. A promising solution to this challenge is the utilization of model editing methods to update the knowledge in an efficient manner. However, the majority of existing model editing techniques are limited to monolingual frameworks, thus failing to address the crucial issue of cross-lingual knowledge synchronization for multilingual models. To tackle this problem, we propose a simple yet effective method that trains multilingual patch neuron to store cross-lingual knowledge. It can be easily adapted to existing approaches to enhance their cross-lingual editing capabilities. To evaluate our method, we conduct experiments using both the XNLI dataset and a self-constructed XFEVER dataset. Experimental results demonstrate that our proposed method achieves improved performance in cross-lingual editing tasks without requiring excessive modifications to the original methodology, thereby showcasing its user-friendly characteristics. Codes will be released soon.
Authors: Aniello Panariello, Gianluca Mancusi, Fedy Haj Ali, Angelo Porrello, Simone Calderara, Rita Cucchiara
Accurate per-object distance estimation is crucial in safety-critical applications such as autonomous driving, surveillance, and robotics. Existing approaches rely on two scales: local information (i.e., the bounding box proportions) or global information, which encodes the semantics of the scene as well as the spatial relations with neighboring objects. However, these approaches may struggle with long-range objects and in the presence of strong occlusions or unusual visual patterns. In this respect, our work aims to strengthen both local and global cues. Our architecture -- named DistFormer -- builds upon three major components acting jointly: i) a robust context encoder extracting fine-grained per-object representations; ii) a masked encoder-decoder module exploiting self-supervision to promote the learning of useful per-object features; iii) a global refinement module that aggregates object representations and computes a joint, spatially-consistent estimation. To evaluate the effectiveness of DistFormer, we conduct experiments on the standard KITTI dataset and the large-scale NuScenes and MOTSynth datasets. Such datasets cover various indoor/outdoor environments, changing weather conditions, appearances, and camera viewpoints. Our comprehensive analysis shows that DistFormer outperforms existing methods. Moreover, we further delve into its generalization capabilities, showing its regularization benefits in zero-shot synth-to-real transfer.
Authors: Ali Falahati, Mohammad Karim Safavi, Ardavan Elahi, Farhad Pakdaman, Moncef Gabbouj
Providing high-quality video with efficient bitrate is a main challenge in video industry. The traditional one-size-fits-all scheme for bitrate ladders is inefficient and reaching the best content-aware decision computationally impractical due to extensive encodings required. To mitigate this, we propose a bitrate and complexity efficient bitrate ladder prediction method using transfer learning and spatio-temporal features. We propose: (1) using feature maps from well-known pre-trained DNNs to predict rate-quality behavior with limited training data; and (2) improving highest quality rung efficiency by predicting minimum bitrate for top quality and using it for the top rung. The method tested on 102 video scenes demonstrates 94.1% reduction in complexity versus brute-force at 1.71% BD-Rate expense. Additionally, transfer learning was thoroughly studied through four networks and ablation studies.
Authors: Zeju Li, Chao Zhang, Xiaoyan Wang, Ruilong Ren, Yifan Xu, Ruifei Ma, Xiangde Liu
The remarkable potential of multi-modal large language models (MLLMs) in comprehending both vision and language information has been widely acknowledged. However, the scarcity of 3D scenes-language pairs in comparison to their 2D counterparts, coupled with the inadequacy of existing approaches in understanding of 3D scenes by LLMs, poses a significant challenge. In response, we collect and construct an extensive dataset comprising 75K instruction-response pairs tailored for 3D scenes. This dataset addresses tasks related to 3D VQA, 3D grounding, and 3D conversation. To further enhance the integration of 3D spatial information into LLMs, we introduce a novel and efficient prompt tuning paradigm, 3DMIT. This paradigm eliminates the alignment stage between 3D scenes and language and extends the instruction prompt with the 3D modality information including the entire scene and segmented objects. We evaluate the effectiveness of our method across diverse tasks in the 3D scene domain and find that our approach serves as a strategic means to enrich LLMs' comprehension of the 3D world. Our code is available at https://github.com/staymylove/3DMIT.
Authors: Tongyan Hua, Haotian Bai, Zidong Cao, Ming Liu, Dacheng Tao, Lin Wang
In this paper, we introduce Hi-Map, a novel monocular dense mapping approach based on Neural Radiance Field (NeRF). Hi-Map is exceptional in its capacity to achieve efficient and high-fidelity mapping using only posed RGB inputs. Our method eliminates the need for external depth priors derived from e.g., a depth estimation model. Our key idea is to represent the scene as a hierarchical feature grid that encodes the radiance and then factorizes it into feature planes and vectors. As such, the scene representation becomes simpler and more generalizable for fast and smooth convergence on new observations. This allows for efficient computation while alleviating noise patterns by reducing the complexity of the scene representation. Buttressed by the hierarchical factorized representation, we leverage the Sign Distance Field (SDF) as a proxy of rendering for inferring the volume density, demonstrating high mapping fidelity. Moreover, we introduce a dual-path encoding strategy to strengthen the photometric cues and further boost the mapping quality, especially for the distant and textureless regions. Extensive experiments demonstrate our method's superiority in geometric and textural accuracy over the state-of-the-art NeRF-based monocular mapping methods.
Authors: Matheus Souza, Wolfgang Heidrich
Image signal processors (ISPs) are historically grown legacy software systems for reconstructing color images from noisy raw sensor measurements. Each smartphone manufacturer has developed its ISPs with its own characteristic heuristics for improving the color rendition, for example, skin tones and other visually essential colors. The recent interest in replacing the historically grown ISP systems with deep-learned pipelines to match DSLR's image quality improves structural features in the image. However, these works ignore the superior color processing based on semantic scene analysis that distinguishes mobile phone ISPs from DSLRs. Here, we present MetaISP, a single model designed to learn how to translate between the color and local contrast characteristics of different devices. MetaISP takes the RAW image from device A as input and translates it to RGB images that inherit the appearance characteristics of devices A, B, and C. We achieve this result by employing a lightweight deep learning technique that conditions its output appearance based on the device of interest. In this approach, we leverage novel attention mechanisms inspired by cross-covariance to learn global scene semantics. Additionally, we use the metadata that typically accompanies RAW images and estimate scene illuminants when they are unavailable.
Authors: Yupei Lin, Xiaoyu Xian, Yukai Shi, Liang Lin
Recently, text-to-image diffusion models become a new paradigm in image processing fields, including content generation, image restoration and image-to-image translation. Given a target prompt, Denoising Diffusion Probabilistic Models (DDPM) are able to generate realistic yet eligible images. With this appealing property, the image translation task has the potential to be free from target image samples for supervision. By using a target text prompt for domain adaption, the diffusion model is able to implement zero-shot image-to-image translation advantageously. However, the sampling and inversion processes of DDPM are stochastic, and thus the inversion process often fail to reconstruct the input content. Specifically, the displacement effect will gradually accumulated during the diffusion and inversion processes, which led to the reconstructed results deviating from the source domain. To make reconstruction explicit, we propose a prompt redescription strategy to realize a mirror effect between the source and reconstructed image in the diffusion model (MirrorDiffusion). More specifically, a prompt redescription mechanism is investigated to align the text prompts with latent code at each time step of the Denoising Diffusion Implicit Models (DDIM) inversion to pursue a structure-preserving reconstruction. With the revised DDIM inversion, MirrorDiffusion is able to realize accurate zero-shot image translation by editing optimized text prompts and latent code. Extensive experiments demonstrate that MirrorDiffusion achieves superior performance over the state-of-the-art methods on zero-shot image translation benchmarks by clear margins and practical model stability.
Authors: Wenqi Ji, Fang liu, Xinxin Du, Niqi Liu, Chao Zhou, Mingjin Yu, Guozhen Zhao, Yong-Jin Liu
Interpersonal relationship quality is pivotal in social and occupational contexts. Existing analysis of interpersonal relationships mostly rely on subjective self-reports, whereas objective quantification remains challenging. In this paper, we propose a novel social relationship analysis framework using spatio-temporal patterns derived from dyadic EEG signals, which can be applied to quantitatively measure team cooperation in corporate team building, and evaluate interpersonal dynamics between therapists and patients in psychiatric therapy. First, we constructed a dyadic-EEG dataset from 72 pairs of participants with two relationships (stranger or friend) when watching emotional videos simultaneously. Then we proposed a deep neural network on dyadic-subject EEG signals, in which we combine the dynamic graph convolutional neural network for characterizing the interpersonal relationships among the EEG channels and 1-dimension convolution for extracting the information from the time sequence. To obtain the feature vectors from two EEG recordings that well represent the relationship of two subjects, we integrate deep canonical correlation analysis and triplet loss for training the network. Experimental results show that the social relationship type (stranger or friend) between two individuals can be effectively identified through their EEG data.
Authors: Shuhao Chen, Yulong Zhang, Weisen Jiang, Jiangang Lu, Yu Zhang
Recent advances achieved by deep learning models rely on the independent and identically distributed assumption, hindering their applications in real-world scenarios with domain shifts. To address the above issues, cross-domain learning aims at extracting domain-invariant knowledge to reduce the domain shift between training and testing data. However, in visual cross-domain learning, traditional methods concentrate solely on the image modality, neglecting the use of the text modality to alleviate the domain shift. In this work, we propose Large Language models as Visual cross-dOmain learners (LLaVO). LLaVO uses vision-language models to convert images into detailed textual descriptions. A large language model is then finetuned on textual descriptions of the source/target domain generated by a designed instruction template. Extensive experimental results on various cross-domain tasks under the domain generalization and unsupervised domain adaptation settings have demonstrated the effectiveness of the proposed method.
Authors: Mengfei Li, Ming Lu, Xiaofang Li, Shanghang Zhang
Recent work on Neural Radiance Fields (NeRF) exploits multi-view 3D consistency, achieving impressive results in 3D scene modeling and high-fidelity novel-view synthesis. However, there are limitations. First, existing methods assume enough high-quality images are available for training the NeRF model, ignoring real-world image degradation. Second, previous methods struggle with ambiguity in the training set due to unmodeled inconsistencies among different views. In this work, we present RustNeRF for real-world high-quality NeRF. To improve NeRF's robustness under real-world inputs, we train a 3D-aware preprocessing network that incorporates real-world degradation modeling. We propose a novel implicit multi-view guidance to address information loss during image degradation and restoration. Extensive experiments demonstrate RustNeRF's advantages over existing approaches under real-world degradation. The code will be released.
Authors: Haritha Thilakarathne, Aiden Nibali, Zhen He, Stuart Morgan
Group activity recognition in video is a complex task due to the need for a model to recognise the actions of all individuals in the video and their complex interactions. Recent studies propose that optimal performance is achieved by individually tracking each person and subsequently inputting the sequence of poses or cropped images/optical flow into a model. This helps the model to recognise what actions each person is performing before they are merged to arrive at the group action class. However, all previous models are highly reliant on high quality tracking and have only been evaluated using ground truth tracking information. In practice it is almost impossible to achieve highly reliable tracking information for all individuals in a group activity video. We introduce an innovative deep learning-based group activity recognition approach called Rendered Pose based Group Activity Recognition System (RePGARS) which is designed to be tolerant of unreliable tracking and pose information. Experimental results confirm that RePGARS outperforms all existing group activity recognition algorithms tested which do not use ground truth detection and tracking information.
Authors: Andrew Gerstenslager, Jomol Lewis, Liam McKenna, Poorva Patel
This paper explores the application of CNN-DNN network fusion to construct a robot navigation controller within a simulated environment. The simulated environment is constructed to model a subterranean rescue situation, such that an autonomous agent is tasked with finding a goal within an unknown cavernous system. Imitation learning is used to train the control algorithm to use LiDAR and camera data to navigate the space and find the goal. The trained model is then tested for robustness using Monte-Carlo.
Authors: Isaiah Lahr, Saghir Alfasly, Peyman Nejat, Jibran Khan, Luke Kottom, Vaishnavi Kumbhar, Areej Alsaafin, Abubakr Shafique, Sobhan Hemati, Ghazal Alabtah, Nneka Comfere, Dennis Murphee, Aaron Mangold, Saba Yasir, Chady Meroueh, Lisa Boardman, Vijay H. Shah, Joaquin J. Garcia, H.R. Tizhoosh
Searching for similar images in archives of histology and histopathology images is a crucial task that may aid in patient matching for various purposes, ranging from triaging and diagnosis to prognosis and prediction. Whole slide images (WSIs) are highly detailed digital representations of tissue specimens mounted on glass slides. Matching WSI to WSI can serve as the critical method for patient matching. In this paper, we report extensive analysis and validation of four search methods bag of visual words (BoVW), Yottixel, SISH, RetCCL, and some of their potential variants. We analyze their algorithms and structures and assess their performance. For this evaluation, we utilized four internal datasets ($1269$ patients) and three public datasets ($1207$ patients), totaling more than $200,000$ patches from $38$ different classes/subtypes across five primary sites. Certain search engines, for example, BoVW, exhibit notable efficiency and speed but suffer from low accuracy. Conversely, search engines like Yottixel demonstrate efficiency and speed, providing moderately accurate results. Recent proposals, including SISH, display inefficiency and yield inconsistent outcomes, while alternatives like RetCCL prove inadequate in both accuracy and efficiency. Further research is imperative to address the dual aspects of accuracy and minimal storage requirements in histopathological image search.
Authors: Walid Guettala, Ali Sayah, Laid Kahloul, Ahmed Tibermacine
One of the most important problems in computer vision and remote sensing is object detection, which identifies particular categories of diverse things in pictures. Two crucial data sources for public security are the thermal infrared (TIR) remote sensing multi-scenario photos and videos produced by unmanned aerial vehicles (UAVs). Due to the small scale of the target, complex scene information, low resolution relative to the viewable videos, and dearth of publicly available labeled datasets and training models, their object detection procedure is still difficult. A UAV TIR object detection framework for pictures and videos is suggested in this study. The Forward-looking Infrared (FLIR) cameras used to gather ground-based TIR photos and videos are used to create the ``You Only Look Once'' (YOLO) model, which is based on CNN architecture. Results indicated that in the validating task, detecting human object had an average precision at IOU (Intersection over Union) = 0.5, which was 72.5\%, using YOLOv7 (YOLO version 7) state of the art model \cite{1}, while the detection speed around 161 frames per second (FPS/second). The usefulness of the YOLO architecture is demonstrated in the application, which evaluates the cross-detection performance of people in UAV TIR videos under a YOLOv7 model in terms of the various UAVs' observation angles. The qualitative and quantitative evaluation of object detection from TIR pictures and videos using deep-learning models is supported favorably by this work.
Authors: Christian Benz, Volker Rodehorst
For effective structural damage assessment, the instances of damages need to be localized in the world of a 3D model. Due to a lack of data, the detection of structural anomalies can currently not be directly learned and performed in 3D space. In this work, a three-stage approach is presented, which uses the good performance of detection models on image level to segment instances of anomalies in the 3D space. In the detection stage, semantic segmentation predictions are produced on image level. The mapping stage transfers the image-level prediction onto the respective point cloud. In the extraction stage, 3D anomaly instances are extracted from the segmented point cloud. Cloud contraction is used to transform cracks into their medial axis representation. For areal anomalies the bounding polygon is extracted by means of alpha shapes. The approach covers the classes crack, spalling, and corrosion and the three image-level segmentation models TopoCrack, nnU-Net, and DetectionHMA are compared. Granted a localization tolerance of 4cm, IoUs of over 90% can be achieved for crack and corrosion and 41% for spalling, which appears to be a specifically challenging class. Detection on instance-level measured in AP is about 45% for crack and spalling and 73% for corrosion.
Authors: Seyed Mohammad Hossein Hashemi, Leila Safari, Amirhossein Dadashzade Taromi
In the field of medical sciences, reliable detection and classification of brain tumors from images remains a formidable challenge due to the rarity of tumors within the population of patients. Therefore, the ability to detect tumors in anomaly scenarios is paramount for ensuring timely interventions and improved patient outcomes. This study addresses the issue by leveraging deep learning (DL) techniques to detect and classify brain tumors in challenging situations. The curated data set from the National Brain Mapping Lab (NBML) comprises 81 patients, including 30 Tumor cases and 51 Normal cases. The detection and classification pipelines are separated into two consecutive tasks. The detection phase involved comprehensive data analysis and pre-processing to modify the number of image samples and the number of patients of each class to anomaly distribution (9 Normal per 1 Tumor) to comply with real world scenarios. Next, in addition to common evaluation metrics for the testing, we employed a novel performance evaluation method called Patient to Patient (PTP), focusing on the realistic evaluation of the model. In the detection phase, we fine-tuned a YOLOv8n detection model to detect the tumor region. Subsequent testing and evaluation yielded competitive performance both in Common Evaluation Metrics and PTP metrics. Furthermore, using the Data Efficient Image Transformer (DeiT) module, we distilled a Vision Transformer (ViT) model from a fine-tuned ResNet152 as a teacher in the classification phase. This approach demonstrates promising strides in reliable tumor detection and classification, offering potential advancements in tumor diagnosis for real-world medical imaging scenarios.
Authors: Arjun Bhalla, Daniel Levenson, Jan Bernhard, Anton Abilov
This work investigates how hierarchically structured data can help neural networks learn conceptual representations of cathedrals. The underlying WikiScenes dataset provides a spatially organized hierarchical structure of cathedral components. We propose a novel hierarchical contrastive training approach that leverages a triplet margin loss to represent the data's spatial hierarchy in the encoder's latent space. As such, the proposed approach investigates if the dataset structure provides valuable information for self-supervised learning. We apply t-SNE to visualize the resultant latent space and evaluate the proposed approach by comparing it with other dataset-specific contrastive learning methods using a common downstream classification task. The proposed method outperforms the comparable weakly-supervised and baseline methods. Our findings suggest that dataset structure is a valuable modality for weakly-supervised learning.
Authors: Qian Gong, Chengzhu Zhang, Xin Liang, Viktor Reshniak, Jieyang Chen, Anand Rangarajan, Sanjay Ranka, Nicolas Vidal, Lipeng Wan, Paul Ullrich, Norbert Podhorszki, Robert Jacob, Scott Klasky
Scientific discoveries are increasingly constrained by limited storage space and I/O capacities. For time-series simulations and experiments, their data often need to be decimated over timesteps to accommodate storage and I/O limitations. In this paper, we propose a technique that addresses storage costs while improving post-analysis accuracy through spatiotemporal adaptive, error-controlled lossy compression. We investigate the trade-off between data precision and temporal output rates, revealing that reducing data precision and increasing timestep frequency lead to more accurate analysis outcomes. Additionally, we integrate spatiotemporal feature detection with data compression and demonstrate that performing adaptive error-bounded compression in higher dimensional space enables greater compression ratios, leveraging the error propagation theory of a transformation-based compressor.
To evaluate our approach, we conduct experiments using the well-known E3SM climate simulation code and apply our method to compress variables used for cyclone tracking. Our results show a significant reduction in storage size while enhancing the quality of cyclone tracking analysis, both quantitatively and qualitatively, in comparison to the prevalent timestep decimation approach. Compared to three state-of-the-art lossy compressors lacking feature preservation capabilities, our adaptive compression framework improves perfectly matched cases in TC tracking by 26.4-51.3% at medium compression ratios and by 77.3-571.1% at large compression ratios, with a merely 5-11% computational overhead.
Authors: Kaiming Fu, Tong Lei, Maryia Halubok, Brian N. Bailey
The accurate identification of walnuts within orchards brings forth a plethora of advantages, profoundly amplifying the efficiency and productivity of walnut orchard management. Nevertheless, the unique characteristics of walnut trees, characterized by their closely resembling shapes, colors, and textures between the walnuts and leaves, present a formidable challenge in precisely distinguishing between them during the annotation process. In this study, we present a novel approach to improve walnut detection efficiency, utilizing YOLOv5 trained on an enriched image set that incorporates both real and synthetic RGB and NIR images. Our analysis comparing results from our original and augmented datasets shows clear improvements in detection when using the synthetic images.
Authors: Dheeraj Vajjarapu
This paper introduces the CowStallNumbers dataset, a collection of images extracted from videos focusing on cow teats, designed to advance the field of cow stall number detection. The dataset comprises 1042 training images and 261 test images, featuring stall numbers ranging from 0 to 60. To enhance the dataset, we performed fine-tuning on a YOLO model and applied data augmentation techniques, including random crop, center crop, and random rotation. The experimental outcomes demonstrate a notable 95.4\% accuracy in recognizing stall numbers.
Authors: Anji Liu, Mathias Niepert, Guy Van den Broeck
Diffusion models are the current state of the art for generating photorealistic images. Controlling the sampling process for constrained image generation tasks such as inpainting, however, remains challenging since exact conditioning on such constraints is intractable. While existing methods use various techniques to approximate the constrained posterior, this paper proposes to exploit the ability of Tractable Probabilistic Models (TPMs) to exactly and efficiently compute the constrained posterior, and to leverage this signal to steer the denoising process of diffusion models. Specifically, this paper adopts a class of expressive TPMs termed Probabilistic Circuits (PCs). Building upon prior advances, we further scale up PCs and make them capable of guiding the image generation process of diffusion models. Empirical results suggest that our approach can consistently improve the overall quality and semantic coherence of inpainted images across three natural image datasets (i.e., CelebA-HQ, ImageNet, and LSUN) with only ~10% additional computational overhead brought by the TPM. Further, with the help of an image encoder and decoder, our method can readily accept semantic constraints on specific regions of the image, which opens up the potential for more controlled image generation tasks. In addition to proposing a new framework for constrained image generation, this paper highlights the benefit of more tractable models and motivates the development of expressive TPMs.
Authors: Martin Bråtelund
The problem of structure from motion is concerned with recovering 3-dimensional structure of an object from a set of 2-dimensional images. Generally, all information can be uniquely recovered if enough images and image points are provided, but there are certain cases where unique recovery is impossible; these are called critical configurations. In this paper we use an algebraic approach to study the critical configurations for two projective cameras. We show that all critical configurations lie on quadric surfaces, and classify exactly which quadrics constitute a critical configuration. The paper also describes the relation between the different reconstructions when unique reconstruction is impossible.
Authors: Zhize Wu, Xiaofeng Wang, Tong Xu, Xuebin Yang, Le Zou, Lixiang Xu, Thomas Weise
Object recognition from images means to automatically find object(s) of interest and to return their category and location information. Benefiting from research on deep learning, like convolutional neural networks~(CNNs) and generative adversarial networks, the performance in this field has been improved significantly, especially when training and test data are drawn from similar distributions. However, mismatching distributions, i.e., domain shifts, lead to a significant performance drop. In this paper, we build domain-invariant detectors by learning domain classifiers via adversarial training. Based on the previous works that align image and instance level features, we mitigate the domain shift further by introducing a domain adaptation component at the region level within Faster \mbox{R-CNN}. We embed a domain classification network in the region proposal network~(RPN) using adversarial learning. The RPN can now generate accurate region proposals in different domains by effectively aligning the features between them. To mitigate the unstable convergence during the adversarial learning, we introduce a balanced domain classifier as well as a network learning rate adjustment strategy. We conduct comprehensive experiments using four standard datasets. The results demonstrate the effectiveness and robustness of our object detection approach in domain shift scenarios.
Authors: Can Ufuk Ertenli, Ramazan Gokberk Cinbis, Emre Akbas
We present StreamDEQ, a method that aims to infer frame-wise representations on videos with minimal per-frame computation. Conventional deep networks do feature extraction from scratch at each frame in the absence of ad-hoc solutions. We instead aim to build streaming recognition models that can natively exploit temporal smoothness between consecutive video frames. We observe that the recently emerging implicit layer models provide a convenient foundation to construct such models, as they define representations as the fixed-points of shallow networks, which need to be estimated using iterative methods. Our main insight is to distribute the inference iterations over the temporal axis by using the most recent representation as a starting point at each frame. This scheme effectively recycles the recent inference computations and greatly reduces the needed processing time. Through extensive experimental analysis, we show that StreamDEQ is able to recover near-optimal representations in a few frames' time and maintain an up-to-date representation throughout the video duration. Our experiments on video semantic segmentation, video object detection, and human pose estimation in videos show that StreamDEQ achieves on-par accuracy with the baseline while being more than 2-4x faster.
Authors: Ziyan Yang, Kushal Kafle, Franck Dernoncourt, Vicente Ordonez
We propose a margin-based loss for tuning joint vision-language models so that their gradient-based explanations are consistent with region-level annotations provided by humans for relatively smaller grounding datasets. We refer to this objective as Attention Mask Consistency (AMC) and demonstrate that it produces superior visual grounding results than previous methods that rely on using vision-language models to score the outputs of object detectors. Particularly, a model trained with AMC on top of standard vision-language modeling objectives obtains a state-of-the-art accuracy of 86.49% in the Flickr30k visual grounding benchmark, an absolute improvement of 5.38% when compared to the best previous model trained under the same level of supervision. Our approach also performs exceedingly well on established benchmarks for referring expression comprehension where it obtains 80.34% accuracy in the easy test of RefCOCO+, and 64.55% in the difficult split. AMC is effective, easy to implement, and is general as it can be adopted by any vision-language model, and can use any type of region annotations.
Authors: Zitong Yu, Rizhao Cai, Zhi Li, Wenhan Yang, Jingang Shi, Alex C. Kot
Face anti-spoofing (FAS) and face forgery detection play vital roles in securing face biometric systems from presentation attacks (PAs) and vicious digital manipulation (e.g., deepfakes). Despite promising performance upon large-scale data and powerful deep models, the generalization problem of existing approaches is still an open issue. Most of recent approaches focus on 1) unimodal visual appearance or physiological (i.e., remote photoplethysmography (rPPG)) cues; and 2) separated feature representation for FAS or face forgery detection. On one side, unimodal appearance and rPPG features are respectively vulnerable to high-fidelity face 3D mask and video replay attacks, inspiring us to design reliable multi-modal fusion mechanisms for generalized face attack detection. On the other side, there are rich common features across FAS and face forgery detection tasks (e.g., periodic rPPG rhythms and vanilla appearance for bonafides), providing solid evidence to design a joint FAS and face forgery detection system in a multi-task learning fashion. In this paper, we establish the first joint face spoofing and forgery detection benchmark using both visual appearance and physiological rPPG cues. To enhance the rPPG periodicity discrimination, we design a two-branch physiological network using both facial spatio-temporal rPPG signal map and its continuous wavelet transformed counterpart as inputs. To mitigate the modality bias and improve the fusion efficacy, we conduct a weighted batch and layer normalization for both appearance and rPPG features before multi-modal fusion. We find that the generalization capacities of both unimodal (appearance or rPPG) and multi-modal (appearance+rPPG) models can be obviously improved via joint training on these two tasks. We hope this new benchmark will facilitate the future research of both FAS and deepfake detection communities.
Authors: Debapriya Roy, Sanchayan Santra, Diganta Mukherjee, Bhabatosh Chanda
The idea of \textit{Virtual Try-ON} (VTON) benefits e-retailing by giving an user the convenience of trying a clothing at the comfort of their home. In general, most of the existing VTON methods produce inconsistent results when a person posing with his arms folded i.e., bent or crossed, wants to try an outfit. The problem becomes severe in the case of long-sleeved outfits. As then, for crossed arm postures, overlap among different clothing parts might happen. The existing approaches, especially the warping-based methods employing \textit{Thin Plate Spline (TPS)} transform can not tackle such cases. To this end, we attempt a solution approach where the clothing from the source person is segmented into semantically meaningful parts and each part is warped independently to the shape of the person. To address the bending issue, we employ hand-crafted geometric features consistent with human body geometry for warping the source outfit. In addition, we propose two learning-based modules: a synthesizer network and a mask prediction network. All these together attempt to produce a photo-realistic, pose-robust VTON solution without requiring any paired training data. Comparison with some of the benchmark methods clearly establishes the effectiveness of the approach.
Authors: Manas Gupta, Efe Camci, Vishandi Rudy Keneta, Abhishek Vaidyanathan, Ritwik Kanodia, Chuan-Sheng Foo, Wu Min, Lin Jie
Pruning neural networks has become popular in the last decade when it was shown that a large number of weights can be safely removed from modern neural networks without compromising accuracy. Numerous pruning methods have been proposed since, each claiming to be better than prior art, however, at the cost of increasingly complex pruning methodologies. These methodologies include utilizing importance scores, getting feedback through back-propagation or having heuristics-based pruning rules amongst others. In this work, we question whether this pattern of introducing complexity is really necessary to achieve better pruning results. We benchmark these SOTA techniques against a simple pruning baseline, namely, Global Magnitude Pruning (Global MP), that ranks weights in order of their magnitudes and prunes the smallest ones. Surprisingly, we find that vanilla Global MP performs very well against the SOTA techniques. When considering sparsity-accuracy trade-off, Global MP performs better than all SOTA techniques at all sparsity ratios. When considering FLOPs-accuracy trade-off, some SOTA techniques outperform Global MP at lower sparsity ratios, however, Global MP starts performing well at high sparsity ratios and performs very well at extremely high sparsity ratios. Moreover, we find that a common issue that many pruning algorithms run into at high sparsity rates, namely, layer-collapse, can be easily fixed in Global MP. We explore why layer collapse occurs in networks and how it can be mitigated in Global MP by utilizing a technique called Minimum Threshold. We showcase the above findings on various models (WRN-28-8, ResNet-32, ResNet-50, MobileNet-V1 and FastGRNN) and multiple datasets (CIFAR-10, ImageNet and HAR-2). Code is available at https://github.com/manasgupta-1/GlobalMP.
Authors: Ruiyuan Lin, Sheng Liu, Jun Jiang, Shujun Li, Chengqing Li, C.-C. Jay Kuo
Recovering unknown, missing, damaged, distorted, or lost information in DCT coefficients is a common task in multiple applications of digital image processing, including image compression, selective image encryption, and image communication. This paper investigates the recovery of sign bits in DCT coefficients of digital images, by proposing two different approximation methods to solve a mixed integer linear programming (MILP) problem, which is NP-hard in general. One method is a relaxation of the MILP problem to a linear programming (LP) problem, and the other splits the original MILP problem into some smaller MILP problems and an LP problem. We considered how the proposed methods can be applied to JPEG-encoded images and conducted extensive experiments to validate their performances. The experimental results showed that the proposed methods outperformed other existing methods by a substantial margin, both according to objective quality metrics and our subjective evaluation.
Authors: Peter Ebert Christensen, Vésteinn Snæbjarnarson, Andrea Dittadi, Serge Belongie, Sagie Benaim
The robustness of image classifiers is essential to their deployment in the real world. The ability to assess this resilience to manipulations or deviations from the training data is thus crucial. These modifications have traditionally consisted of minimal changes that still manage to fool classifiers, and modern approaches are increasingly robust to them. Semantic manipulations that modify elements of an image in meaningful ways have thus gained traction for this purpose. However, they have primarily been limited to style, color, or attribute changes. While expressive, these manipulations do not make use of the full capabilities of a pretrained generative model. In this work, we aim to bridge this gap. We show how a pretrained image generator can be used to semantically manipulate images in a detailed, diverse, and photorealistic way while still preserving the class of the original image. Inspired by recent GAN-based image inversion methods, we propose a method called Adversarial Pivotal Tuning (APT). Given an image, APT first finds a pivot latent space input that reconstructs the image using a pretrained generator. It then adjusts the generator's weights to create small yet semantic manipulations in order to fool a pretrained classifier. APT preserves the full expressive editing capabilities of the generative model. We demonstrate that APT is capable of a wide range of class-preserving semantic image manipulations that fool a variety of pretrained classifiers. Finally, we show that classifiers that are robust to other benchmarks are not robust to APT manipulations and suggest a method to improve them. Code available at: https://captaine.github.io/apt/
Authors: Xiuding Cai, Yaoyao Zhu, Dong Miao, Linjie Fu, Yu Yao
In an unpaired setting, lacking sufficient content constraints for image-to-image translation (I2I) tasks, GAN-based approaches are usually prone to model collapse. Current solutions can be divided into two categories, reconstruction-based and Siamese network-based. The former requires that the transformed or transforming image can be perfectly converted back to the original image, which is sometimes too strict and limits the generative performance. The latter involves feeding the original and generated images into a feature extractor and then matching their outputs. This is not efficient enough, and a universal feature extractor is not easily available. In this paper, we propose EnCo, a simple but efficient way to maintain the content by constraining the representational similarity in the latent space of patch-level features from the same stage of the \textbf{En}coder and de\textbf{Co}der of the generator. For the similarity function, we use a simple MSE loss instead of contrastive loss, which is currently widely used in I2I tasks. Benefits from the design, EnCo training is extremely efficient, while the features from the encoder produce a more positive effect on the decoding, leading to more satisfying generations. In addition, we rethink the role played by discriminators in sampling patches and propose a discriminative attention-guided (DAG) patch sampling strategy to replace random sampling. DAG is parameter-free and only requires negligible computational overhead, while significantly improving the performance of the model. Extensive experiments on multiple datasets demonstrate the effectiveness and advantages of EnCo, and we achieve multiple state-of-the-art compared to previous methods. Our code is available at https://github.com/XiudingCai/EnCo-pytorch.
Authors: Chuanming Tang, Xiao Wang, Ju Huang, Bo Jiang, Lin Zhu, Jianlin Zhang, Yaowei Wang, Yonghong Tian
Combining the Color and Event cameras (also called Dynamic Vision Sensors, DVS) for robust object tracking is a newly emerging research topic in recent years. Existing color-event tracking framework usually contains multiple scattered modules which may lead to low efficiency and high computational complexity, including feature extraction, fusion, matching, interactive learning, etc. In this paper, we propose a single-stage backbone network for Color-Event Unified Tracking (CEUTrack), which achieves the above functions simultaneously. Given the event points and RGB frames, we first transform the points into voxels and crop the template and search regions for both modalities, respectively. Then, these regions are projected into tokens and parallelly fed into the unified Transformer backbone network. The output features will be fed into a tracking head for target object localization. Our proposed CEUTrack is simple, effective, and efficient, which achieves over 75 FPS and new SOTA performance. To better validate the effectiveness of our model and address the data deficiency of this task, we also propose a generic and large-scale benchmark dataset for color-event tracking, termed COESOT, which contains 90 categories and 1354 video sequences. Additionally, a new evaluation metric named BOC is proposed in our evaluation toolkit to evaluate the prominence with respect to the baseline methods. We hope the newly proposed method, dataset, and evaluation metric provide a better platform for color-event-based tracking. The dataset, toolkit, and source code will be released on: \url{https://github.com/Event-AHU/COESOT}.
Authors: Xu Yan, Chaoda Zheng, Ying Xue, Zhen Li, Shuguang Cui, Dengxin Dai
When using LiDAR semantic segmentation models for safety-critical applications such as autonomous driving, it is essential to understand and improve their robustness with respect to a large range of LiDAR corruptions. In this paper, we aim to comprehensively analyze the robustness of LiDAR semantic segmentation models under various corruptions. To rigorously evaluate the robustness and generalizability of current approaches, we propose a new benchmark called SemanticKITTI-C, which features 16 out-of-domain LiDAR corruptions in three groups, namely adverse weather, measurement noise and cross-device discrepancy. Then, we systematically investigate 11 LiDAR semantic segmentation models, especially spanning different input representations (e.g., point clouds, voxels, projected images, and etc.), network architectures and training schemes. Through this study, we obtain two insights: 1) We find out that the input representation plays a crucial role in robustness. Specifically, under specific corruptions, different representations perform variously. 2) Although state-of-the-art methods on LiDAR semantic segmentation achieve promising results on clean data, they are less robust when dealing with noisy data. Finally, based on the above observations, we design a robust LiDAR segmentation model (RLSeg) which greatly boosts the robustness with simple but effective modifications. It is promising that our benchmark, comprehensive analysis, and observations can boost future research in robust LiDAR semantic segmentation for safety-critical applications.
Authors: Nicolas Nerrienet, Rémy Peyret, Marie Sockeel, Stéphane Sockeel
Generalization is one of the main challenges of computational pathology. Slide preparation heterogeneity and the diversity of scanners lead to poor model performance when used on data from medical centers not seen during training. In order to achieve stain invariance in breast invasive carcinoma patch classification, we implement a stain translation strategy using cycleGANs for unsupervised image-to-image translation. We compare three cycleGAN-based approaches to a baseline classification model obtained without any stain invariance strategy. Two of the proposed approaches use cycleGAN's translations at inference or training in order to build stain-specific classification models. The last method uses them for stain data augmentation during training. This constrains the classification model to learn stain-invariant features. Baseline metrics are set by training and testing the baseline classification model on a reference stain. We assessed performances using three medical centers with H&E and H&E&S staining. Every approach tested in this study improves baseline metrics without needing labels on target stains. The stain augmentation-based approach produced the best results on every stain. Each method's pros and cons are studied and discussed in this paper. However, training highly performing cycleGANs models in itself represents a challenge. In this work, we introduce a systematical method for optimizing cycleGAN training by setting a novel stopping criterion. This method has the benefit of not requiring any visual inspection of cycleGAN results and proves superiority to methods using a predefined number of training epochs. In addition, we also study the minimal amount of data required for cycleGAN training.
Authors: Yuanpeng Tu, Yuxi Li, Boshen Zhang, Liang Liu, Jiangning Zhang, Yabiao Wang, Cai Rong Zhao
Robust autonomous driving requires agents to accurately identify unexpected areas (anomalies) in urban scenes. To this end, some critical issues remain open: how to design advisable metric to measure anomalies, and how to properly generate training samples of anomaly data? Classical effort in anomaly detection usually resorts to pixel-wise uncertainty or sample synthesis, which ignores the contextual information and sometimes requires auxiliary data with fine-grained annotations. On the contrary, in this paper, we exploit the strong context-dependent nature of the segmentation task and design an energy-guided self-supervised framework for anomaly segmentation, which optimizes an anomaly head by maximizing the likelihood of self-generated anomaly pixels. For this purpose, we design two estimators to model anomaly likelihood, one is a task-agnostic binary estimator and the other depicts the likelihood as residual of task-oriented joint energy. Based on the proposed estimators, we devise an adaptive self-supervised training framework, which exploits the contextual reliance and estimated likelihood to refine mask annotations in anomaly areas. We conduct extensive experiments on challenging Fishyscapes and Road Anomaly benchmarks, demonstrating that without any auxiliary data or synthetic models, our method can still achieve comparable performance to supervised competitors. Code is available at https://github.com/yuanpengtu/SLEEG..
Authors: Zhihao Chen, Chuang Niu, Qi Gao, Ge Wang, Hongming Shan
This paper studies 3D low-dose computed tomography (CT) imaging. Although various deep learning methods were developed in this context, typically they focus on 2D images and perform denoising due to low-dose and deblurring for super-resolution separately. Up to date, little work was done for simultaneous in-plane denoising and through-plane deblurring, which is important to obtain high-quality 3D CT images with lower radiation and faster imaging speed. For this task, a straightforward method is to directly train an end-to-end 3D network. However, it demands much more training data and expensive computational costs. Here, we propose to link in-plane and through-plane transformers for simultaneous in-plane denoising and through-plane deblurring, termed as LIT-Former, which can efficiently synergize in-plane and through-plane sub-tasks for 3D CT imaging and enjoy the advantages of both convolution and transformer networks. LIT-Former has two novel designs: efficient multi-head self-attention modules (eMSM) and efficient convolutional feedforward networks (eCFN). First, eMSM integrates in-plane 2D self-attention and through-plane 1D self-attention to efficiently capture global interactions of 3D self-attention, the core unit of transformer networks. Second, eCFN integrates 2D convolution and 1D convolution to extract local information of 3D convolution in the same fashion. As a result, the proposed LIT-Former synergize these two subtasks, significantly reducing the computational complexity as compared to 3D counterparts and enabling rapid convergence. Extensive experimental results on simulated and clinical datasets demonstrate superior performance over state-of-the-art models. The source code is made available at https://github.com/hao1635/LIT-Former.
Authors: Marco D'Alessandro, Alberto Alonso, Enrique Calabrés, Mikel Galar
Few-Shot Class Incremental Learning (FSCIL) is a challenging continual learning task, where limited training examples are available during several learning sessions. To succeed in this task, it is necessary to avoid over-fitting new classes caused by biased distributions in the few-shot training sets. The general approach to address this issue involves enhancing the representational capability of a pre-defined backbone architecture by adding special modules for backward compatibility with older classes. However, this approach has not yet solved the dilemma of ensuring high classification accuracy over time while reducing the gap between the performance obtained on larger training sets and the smaller ones. In this work, we propose an alternative approach called Continual Parameter-Efficient CLIP (CPE-CLIP) to reduce the loss of information between different learning sessions. Instead of adapting additional modules to address information loss, we leverage the vast knowledge acquired by CLIP in large-scale pre-training and its effectiveness in generalizing to new concepts. Our approach is multimodal and parameter-efficient, relying on learnable prompts for both the language and vision encoders to enable transfer learning across sessions. We also introduce prompt regularization to improve performance and prevent forgetting. Our experimental results demonstrate that CPE-CLIP significantly improves FSCIL performance compared to state-of-the-art proposals while also drastically reducing the number of learnable parameters and training costs.
Authors: Haijian Chen, Wendong Zhang, Yunbo Wang, Xiaokang Yang
Masked image modeling is a promising self-supervised learning method for visual data. It is typically built upon image patches with random masks, which largely ignores the variation of information density between them. The question is: Is there a better masking strategy than random sampling and how can we learn it? We empirically study this problem and initially find that introducing object-centric priors in mask sampling can significantly improve the learned representations. Inspired by this observation, we present AutoMAE, a fully differentiable framework that uses Gumbel-Softmax to interlink an adversarially-trained mask generator and a mask-guided image modeling process. In this way, our approach can adaptively find patches with higher information density for different images, and further strike a balance between the information gain obtained from image reconstruction and its practical training difficulty. In our experiments, AutoMAE is shown to provide effective pretraining models on standard self-supervised benchmarks and downstream tasks.
Authors: Hao Zhang, Yeo Keat Ee, Basura Fernando
Visual Abductive Reasoning is an emerging vision-language (VL) topic where the model needs to retrieve/generate a likely textual hypothesis from a visual input (image or its part) using backward reasoning based on commonsense. Unlike in conventional VL retrieval or captioning tasks, where entities of texts appear in the image, in abductive inferences, the relevant facts about inferences are not readily apparent in the input images. Besides, these inferences are causally linked to specific regional visual cues and would change as cues change. Existing works highlight cues utilizing a specific prompt (e.g., colorful prompt). Then, a full fine-tuning of a VL foundation model is launched to tweak its function from perception to deduction. However, the colorful prompt uniformly patchify ``regional hints'' and ``global context'' at the same granularity level and may lose fine-grained visual details crucial for VAR. Meanwhile, full fine-tuning of VLF on limited data would easily be overfitted.
To tackle this, we propose a simple yet effective Region-Prompted Adapter (RPA), a hybrid parameter-efficient fine-tuning method that leverages the strengths of detailed cues and efficient training for the VAR task. RPA~consists of two novel modules: Regional Prompt Generator (RPG) and Adapter$^\textbf{+}$. The prior encodes ``regional visual hints'' and ``global contexts'' into visual prompts separately at fine and coarse-grained levels. The latter extends the vanilla adapters with a new Map Adapter, which modifies the attention map using a trainable low-dim query/key projection. Additionally, we propose a new Dual-Contrastive Loss to regress the visual feature toward features of factual description and plausible hypothesis. Experiments on the Sherlock demonstrate that RPA outperforms previous SOTAs, achieving the 1st rank on leaderboards (Comparison to Human Accuracy: RPA~31.74 vs CPT-CLIP 29.58).
Authors: Sander Riisøen Jyhne, Per-Arne Andersen, Morten Goodwin
This paper presents Contrastive Transformer, a contrastive learning scheme using the Transformer innate patches. Contrastive Transformer enables existing contrastive learning techniques, often used for image classification, to benefit dense downstream prediction tasks such as semantic segmentation. The scheme performs supervised patch-level contrastive learning, selecting the patches based on the ground truth mask, subsequently used for hard-negative and hard-positive sampling. The scheme applies to all vision-transformer architectures, is easy to implement, and introduces minimal additional memory footprint. Additionally, the scheme removes the need for huge batch sizes, as each patch is treated as an image.
We apply and test Contrastive Transformer for the case of aerial image segmentation, known for low-resolution data, large class imbalance, and similar semantic classes. We perform extensive experiments to show the efficacy of the Contrastive Transformer scheme on the ISPRS Potsdam aerial image segmentation dataset. Additionally, we show the generalizability of our scheme by applying it to multiple inherently different Transformer architectures. Ultimately, the results show a consistent increase in mean IoU across all classes.
Authors: Haimei Zhao, Qiming Zhang, Shanshan Zhao, Zhe Chen, Jing Zhang, Dacheng Tao
Multi-view camera-based 3D object detection has become popular due to its low cost, but accurately inferring 3D geometry solely from camera data remains challenging and may lead to inferior performance. Although distilling precise 3D geometry knowledge from LiDAR data could help tackle this challenge, the benefits of LiDAR information could be greatly hindered by the significant modality gap between different sensory modalities. To address this issue, we propose a Simulated multi-modal Distillation (SimDistill) method by carefully crafting the model architecture and distillation strategy. Specifically, we devise multi-modal architectures for both teacher and student models, including a LiDAR-camera fusion-based teacher and a simulated fusion-based student. Owing to the ``identical'' architecture design, the student can mimic the teacher to generate multi-modal features with merely multi-view images as input, where a geometry compensation module is introduced to bridge the modality gap. Furthermore, we propose a comprehensive multi-modal distillation scheme that supports intra-modal, cross-modal, and multi-modal fusion distillation simultaneously in the Bird's-eye-view space. Incorporating them together, our SimDistill can learn better feature representations for 3D object detection while maintaining a cost-effective camera-only deployment. Extensive experiments validate the effectiveness and superiority of SimDistill over state-of-the-art methods, achieving an improvement of 4.8\% mAP and 4.1\% NDS over the baseline detector. The source code will be released at https://github.com/ViTAE-Transformer/SimDistill.
Authors: Juan Terven, Diana Cordova-Esparza
YOLO has become a central real-time object detection system for robotics, driverless cars, and video monitoring applications. We present a comprehensive analysis of YOLO's evolution, examining the innovations and contributions in each iteration from the original YOLO up to YOLOv8, YOLO-NAS, and YOLO with Transformers. We start by describing the standard metrics and postprocessing; then, we discuss the major changes in network architecture and training tricks for each model. Finally, we summarize the essential lessons from YOLO's development and provide a perspective on its future, highlighting potential research directions to enhance real-time object detection systems.
Authors: Akshaya Athwale, Ichrak Shili, Émile Bergeron, Arman Afrasiyabi, Justin Lagüe, Ola Ahmad, Jean-François Lalonde
Wide-angle lenses are commonly used in perception tasks requiring a large field of view. Unfortunately, these lenses produce significant distortions, making conventional models that ignore the distortion effects unable to adapt to wide-angle images. In this paper, we present a novel transformer-based model that automatically adapts to the distortion produced by wide-angle lenses. Our proposed image encoder architecture, dubbed DarSwin, leverages the physical characteristics of such lenses analytically defined by the radial distortion profile. In contrast to conventional transformer-based architectures, DarSwin comprises a radial patch partitioning, a distortion-based sampling technique for creating token embeddings, and an angular position encoding for radial patch merging. Compared to other baselines, DarSwin achieves the best results on different datasets with significant gains when trained on bounded levels of distortions (very low, low, medium, and high) and tested on all, including out-of-distribution distortions. While the base DarSwin architecture requires knowledge of the radial distortion profile, we show it can be combined with a self-calibration network that estimates such a profile from the input image itself, resulting in a completely uncalibrated pipeline. Finally, we also present DarSwin-Unet, which extends DarSwin, to an encoder-decoder architecture suitable for pixel-level tasks. We demonstrate its performance on depth estimation and show through extensive experiments that DarSwin-Unet can perform zero-shot adaptation to unseen distortions of different wide-angle lenses. The code and models are publicly available at https://lvsn.github.io/darswin/
Authors: Bin Fan, Yuchao Dai, Yongduek Seo, Mingyi He
The Normalized Eight-Point algorithm has been widely viewed as the cornerstone in two-view geometry computation, where the seminal Hartley's normalization greatly improves the performance of the direct linear transformation (DLT) algorithm. A natural question is, whether there exists and how to find other normalization methods that may further improve the performance as per each input sample. In this paper, we provide a novel perspective and make two contributions towards this fundamental problem: 1) We revisit the normalized eight-point algorithm and make a theoretical contribution by showing the existence of different and better normalization algorithms; 2) We present a deep convolutional neural network with a self-supervised learning strategy to the normalization. Given eight pairs of correspondences, our network directly predicts the normalization matrices, thus learning to normalize each input sample. Our learning-based normalization module could be integrated with both traditional (e.g., RANSAC) and deep learning framework (affording good interpretability) with minimal efforts. Extensive experiments on both synthetic and real images show the effectiveness of our proposed approach.
Authors: Hanqing Sun, Yanwei Pang, Jiale Cao, Jin Xie, Xuelong Li
Transformers have shown promising progress in various visual object detection tasks, including monocular 2D/3D detection and surround-view 3D detection. More importantly, the attention mechanism in the Transformer model and the image correspondence in binocular stereo are both similarity-based. However, directly applying existing Transformer-based detectors to binocular stereo 3D object detection leads to slow convergence and significant precision drops. We argue that a key cause of this defect is that existing Transformers ignore the stereo-specific image correspondence information. In this paper, we explore the model design of Transformers in binocular 3D object detection, focusing particularly on extracting and encoding the task-specific image correspondence information. To achieve this goal, we present TS3D, a Transformer-based Stereo-aware 3D object detector. In the TS3D, a Disparity-Aware Positional Encoding (DAPE) module is proposed to embed the image correspondence information into stereo features. The correspondence is encoded as normalized sub-pixel-level disparity and is used in conjunction with sinusoidal 2D positional encoding to provide the 3D location information of the scene. To extract enriched multi-scale stereo features, we propose a Stereo Preserving Feature Pyramid Network (SPFPN). The SPFPN is designed to preserve the correspondence information while fusing intra-scale and aggregating cross-scale stereo features. Our proposed TS3D achieves a 41.29% Moderate Car detection average precision on the KITTI test set and takes 88 ms to detect objects from each binocular image pair. It is competitive with advanced counterparts in terms of both precision and inference speed.
Authors: Guang Jun Nicholas Ang, Aritejh Kr Goil, Henryk Chan, Jieyi Jeric Lew, Xin Chun Lee, Raihan Bin Ahmad Mustaffa, Timotius Jason, Ze Ting Woon, Bingquan Shen
In a landscape characterized by heightened connectivity and mobility, coupled with a surge in cardiovascular ailments, the imperative to curtail healthcare expenses through remote monitoring of cardiovascular health has become more pronounced. The accurate detection and classification of cardiac arrhythmias are pivotal for diagnosing individuals with heart irregularities. This study underscores the feasibility of employing electrocardiograms (ECG) measurements in the home environment for real-time arrhythmia detection. Presenting a fresh application for arrhythmia detection, this paper leverages the cutting-edge You-Only-Look-Once (YOLO)v8 algorithm to categorize single-lead ECG signals. We introduce a novel loss-modified YOLOv8 model, fine-tuned on the MIT-BIH arrhythmia dataset, enabling real-time continuous monitoring. The obtained results substantiate the efficacy of our approach, with the model attaining an average accuracy of 99.5% and 0.992 mAP@50, and a rapid detection time of 0.002 seconds on an NVIDIA Tesla V100. Our investigation exemplifies the potential of real-time arrhythmia detection, enabling users to visually interpret the model output within the comfort of their homes. Furthermore, this study lays the groundwork for an extension into a real-time explainable AI (XAI) model capable of deployment in the healthcare sector, thereby significantly advancing the realm of healthcare solutions.
Authors: Yanwen Fang, Jintai Chen, Peng-Tao Jiang, Chao Li, Yifeng Geng, Eddy K. F. Lam, Guodong Li
Multi-person motion prediction is a challenging task, especially for real-world scenarios of highly interacted persons. Most previous works have been devoted to studying the case of weak interactions (e.g., walking together), in which typically forecasting each human pose in isolation can still achieve good performances. This paper focuses on collaborative motion prediction for multiple persons with extreme motions and attempts to explore the relationships between the highly interactive persons' pose trajectories. Specifically, a novel cross-query attention (XQA) module is proposed to bilaterally learn the cross-dependencies between the two pose sequences tailored for this situation. A proxy unit is additionally introduced to bridge the involved persons, which cooperates with our proposed XQA module and subtly controls the bidirectional spatial information flows. These designs are then integrated into a Transformer-based architecture and the resulting model is called Proxy-bridged Game Transformer (PGformer) for multi-person interactive motion prediction. Its effectiveness has been evaluated on the challenging ExPI dataset, which involves highly interactive actions. Our PGformer consistently outperforms the state-of-the-art methods in both short- and long-term predictions by a large margin. Besides, our approach can also be compatible with the weakly interacted CMU-Mocap and MuPoTS-3D datasets and extended to the case of more than 2 individuals with encouraging results.
Authors: Pengzhen Ren, Kaidong Zhang, Hetao Zheng, Zixuan Li, Yuhang Wen, Fengda Zhu, Mas Ma, Xiaodan Liang
The research on embodied AI has greatly promoted the development of robot manipulation. However, it still faces significant challenges in various aspects such as benchmark construction, multi-modal perception and decision-making, and physical execution. Previous robot manipulation simulators were primarily designed to enrich manipulation types and types of objects while neglecting the balance between physical manipulation and language instruction complexity in multi-modal environments. This paper proposes a new robot manipulation simulator and builds a comprehensive and systematic robot manipulation benchmark with progressive reasoning tasks called SeaWave (i.e., a progressive reasoning benchmark). It provides a standard test platform for embedded AI agents in a multi-modal environment, which can evaluate and execute four levels of human natural language instructions at the same time.
Previous world model-based robot manipulation work lacked research on the perception and decision-making of complex instructions in multi-modal environments. To this end, we propose a new world model tailored for cross-modal robot manipulation called DamWorld. Specifically, DamWorld takes the current visual scene and predicted execution actions based on natural language instructions as input, and uses the next action frame to supervise the output of the world model to force the model to learn robot manipulation consistent with world knowledge. Compared with the renowned baselines (e.g., RT-1), our DamWorld improves the manipulation success rate by 5.6% on average on four levels of progressive reasoning tasks. It is worth noting that on the most challenging level 4 manipulation task, DamWorld still improved by 9.0% compared to prior works.
Authors: Xiao Liu, Guangyi Chen, Yansong Tang, Guangrun Wang, Xiao-Ping Zhang, Ser-Nam Lim
Composing simple elements into complex concepts is crucial yet challenging, especially for 3D action generation. Existing methods largely rely on extensive neural language annotations to discern composable latent semantics, a process that is often costly and labor-intensive. In this study, we introduce a novel framework to generate compositional actions without reliance on language auxiliaries. Our approach consists of three main components: Action Coupling, Conditional Action Generation, and Decoupling Refinement. Action Coupling utilizes an energy model to extract the attention masks of each sub-action, subsequently integrating two actions using these attentions to generate pseudo-training examples. Then, we employ a conditional generative model, CVAE, to learn a latent space, facilitating the diverse generation. Finally, we propose Decoupling Refinement, which leverages a self-supervised pre-trained model MAE to ensure semantic consistency between the sub-actions and compositional actions. This refinement process involves rendering generated 3D actions into 2D space, decoupling these images into two sub-segments, using the MAE model to restore the complete image from sub-segments, and constraining the recovered images to match images rendered from raw sub-actions. Due to the lack of existing datasets containing both sub-actions and compositional actions, we created two new datasets, named HumanAct-C and UESTC-C, and present a corresponding evaluation metric. Both qualitative and quantitative assessments are conducted to show our efficacy.
Authors: Zhe Huang, Ruijie Jiang, Shuchin Aeron, Michael C. Hughes
In many medical image classification problems, labeled data is scarce while unlabeled data is more available. Semi-supervised learning and self-supervised learning are two different research directions that can improve accuracy by learning from extra unlabeled data. Recent methods from both directions have reported significant gains on traditional benchmarks. Yet past benchmarks do not focus on medical tasks and rarely compare self- and semi- methods together on equal footing. Furthermore, past benchmarks often handle hyperparameter tuning suboptimally. First, they may not tune hyperparameters at all, leading to underfitting. Second, when tuning does occur, it often unrealistically uses a labeled validation set much larger than the train set. Both cases make previously published rankings of methods difficult to translate to practical settings. This study contributes a systematic evaluation of self- and semi- methods with a unified experimental protocol intended to guide a practitioner with scarce overall labeled data and a limited compute budget. We answer two key questions: Can hyperparameter tuning be effective with realistic-sized validation sets? If so, when all methods are tuned well, which self- or semi-supervised methods reach the best accuracy? Our study compares 13 representative semi- and self-supervised methods to strong labeled-set-only baselines on 4 medical datasets. From 20000+ total GPU hours of computation, we provide valuable best practices to resource-constrained, results-focused practitioners.
Authors: Seyed Rohollah Hosseyni, Sanaz Seyedin, Hasan Taheri
Understanding the relationship between different parts of an image is crucial in a variety of applications, including object recognition, scene understanding, and image classification. Despite the fact that Convolutional Neural Networks (CNNs) have demonstrated impressive results in classifying and detecting objects, they lack the capability to extract the relationship between different parts of an image, which is a crucial factor in Human Action Recognition (HAR). To address this problem, this paper proposes a new module that functions like a convolutional layer that uses Vision Transformer (ViT). In the proposed model, the Vision Transformer can complement a convolutional neural network in a variety of tasks by helping it to effectively extract the relationship among various parts of an image. It is shown that the proposed model, compared to a simple CNN, can extract meaningful parts of an image and suppress the misleading parts. The proposed model has been evaluated on the Stanford40 and PASCAL VOC 2012 action datasets and has achieved 95.5% mean Average Precision (mAP) and 91.5% mAP results, respectively, which are promising compared to other state-of-the-art methods.
Authors: Wei Jiang, Jiayu Yang, Yongqi Zhai, Feng Gao, Ronggang Wang
Recently, learned image compression has achieved impressive performance. The entropy model, which estimates the distribution of the latent representation, plays a crucial role in enhancing rate-distortion performance. However, existing global context modules rely on computationally intensive quadratic complexity computations to capture global correlations. This quadratic complexity imposes limitations on the potential of high-resolution image coding. Moreover, effectively capturing local, global, and channel-wise contexts with acceptable even linear complexity within a single entropy model remains a challenge. To address these limitations, we propose the Linear Complexity Attention-based Multi-Reference Entropy Model (MEM++). MEM++ effectively captures the diverse range of correlations inherent in the latent representation. Specifically, the latent representation is first divided into multiple slices. When compressing a particular slice, the previously compressed slices serve as its channel-wise contexts. To capture local contexts without sacrificing performance, we introduce a novel checkerboard attention module. Additionally, to capture global contexts, we propose the linear complexity attention-based global correlations capturing by leveraging the decomposition of the softmax operation. The attention map of the previously decoded slice is implicitly computed and employed to predict global correlations in the current slice. Based on MEM++, we propose image compression model MLIC++. Extensive experimental evaluations demonstrate that our MLIC++ achieves state-of-the-art performance, reducing BD-rate by 13.39% on the Kodak dataset compared to VTM-17.0 in PSNR. Furthermore, MLIC++ exhibits linear GPU memory consumption with resolution, making it highly suitable for high-resolution image coding. Code and pre-trained models are available at https://github.com/JiangWeibeta/MLIC.
Authors: Ben Chen, Xuechao Zou, Yu Zhang, Jiayu Li, Kai Li, Junliang Xing, Pin Tao
Lake extraction from remote sensing images is challenging due to the complex lake shapes and inherent data noises. Existing methods suffer from blurred segmentation boundaries and poor foreground modeling. This paper proposes a hybrid CNN-Transformer architecture, called LEFormer, for accurate lake extraction. LEFormer contains three main modules: CNN encoder, Transformer encoder, and cross-encoder fusion. The CNN encoder effectively recovers local spatial information and improves fine-scale details. Simultaneously, the Transformer encoder captures long-range dependencies between sequences of any length, allowing them to obtain global features and context information. The cross-encoder fusion module integrates the local and global features to improve mask prediction. Experimental results show that LEFormer consistently achieves state-of-the-art performance and efficiency on the Surface Water and the Qinghai-Tibet Plateau Lake datasets. Specifically, LEFormer achieves 90.86% and 97.42% mIoU on two datasets with a parameter count of 3.61M, respectively, while being 20 minor than the previous best lake extraction method. The source code is available at https://github.com/BastianChen/LEFormer.
Authors: Qiu Zhou, Jinming Cao, Hanchao Leng, Yifang Yin, Yu Kun, Roger Zimmermann
In the field of autonomous driving, accurate and comprehensive perception of the 3D environment is crucial. Bird's Eye View (BEV) based methods have emerged as a promising solution for 3D object detection using multi-view images as input. However, existing 3D object detection methods often ignore the physical context in the environment, such as sidewalk and vegetation, resulting in sub-optimal performance. In this paper, we propose a novel approach called SOGDet (Semantic-Occupancy Guided Multi-view 3D Object Detection), that leverages a 3D semantic-occupancy branch to improve the accuracy of 3D object detection. In particular, the physical context modeled by semantic occupancy helps the detector to perceive the scenes in a more holistic view. Our SOGDet is flexible to use and can be seamlessly integrated with most existing BEV-based methods. To evaluate its effectiveness, we apply this approach to several state-of-the-art baselines and conduct extensive experiments on the exclusive nuScenes dataset. Our results show that SOGDet consistently enhance the performance of three baseline methods in terms of nuScenes Detection Score (NDS) and mean Average Precision (mAP). This indicates that the combination of 3D object detection and 3D semantic occupancy leads to a more comprehensive perception of the 3D environment, thereby aiding build more robust autonomous driving systems. The codes are available at: https://github.com/zhouqiu/SOGDet.
Authors: Chaofan Ma, Yuhuan Yang, Chen Ju, Fei Zhang, Ya Zhang, Yanfeng Wang
Open-vocabulary semantic segmentation is a challenging task that requires segmenting novel object categories at inference time. Recent studies have explored vision-language pre-training to handle this task, but suffer from unrealistic assumptions in practical scenarios, i.e., low-quality textual category names. For example, this paradigm assumes that new textual categories will be accurately and completely provided, and exist in lexicons during pre-training. However, exceptions often happen when encountering ambiguity for brief or incomplete names, new words that are not present in the pre-trained lexicons, and difficult-to-describe categories for users. To address these issues, this work proposes a novel attribute decomposition-aggregation framework, AttrSeg, inspired by human cognition in understanding new concepts. Specifically, in the decomposition stage, we decouple class names into diverse attribute descriptions to complement semantic contexts from multiple perspectives. Two attribute construction strategies are designed: using large language models for common categories, and involving manually labeling for human-invented categories. In the aggregation stage, we group diverse attributes into an integrated global description, to form a discriminative classifier that distinguishes the target object from others. One hierarchical aggregation architecture is further proposed to achieve multi-level aggregations, leveraging the meticulously designed clustering module. The final results are obtained by computing the similarity between aggregated attributes and images embeddings. To evaluate the effectiveness, we annotate three types of datasets with attribute descriptions, and conduct extensive experiments and ablation studies. The results show the superior performance of attribute decomposition-aggregation.
Authors: Wenxuan Zhang, Xuechao Zou, Li Wu, Xiaoying Wang, Jianqiang Huang, Junliang Xing
Spatiotemporal prediction aims to generate future sequences by paradigms learned from historical contexts. It is essential in numerous domains, such as traffic flow prediction and weather forecasting. Recently, research in this field has been predominantly driven by deep neural networks based on autoencoder architectures. However, existing methods commonly adopt autoencoder architectures with identical receptive field sizes. To address this issue, we propose an Asymmetric Receptive Field Autoencoder (ARFA) model, which introduces corresponding sizes of receptive field modules tailored to the distinct functionalities of the encoder and decoder. In the encoder, we present a large kernel module for global spatiotemporal feature extraction. In the decoder, we develop a small kernel module for local spatiotemporal information reconstruction. Experimental results demonstrate that ARFA consistently achieves state-of-the-art performance on popular datasets. Additionally, we construct the RainBench, a large-scale radar echo dataset for precipitation prediction, to address the scarcity of meteorological data in the domain.
Authors: Harini S I, Somesh Singh, Yaman K Singla, Aanisha Bhattacharyya, Veeky Baths, Changyou Chen, Rajiv Ratn Shah, Balaji Krishnamurthy
Marketers spend billions of dollars on advertisements but to what end? At the time of purchase, if customers cannot recognize the brand for which they saw an ad, the money spent on the ad is essentially wasted. Despite its importance in marketing, until now, there has been no study on the memorability of ads in the ML literature. Most studies have been conducted on short-term recall (<5 mins) on specific content types like object and action videos. On the other hand, the advertising industry only cares about long-term memorability, and ads are almost always highly multimodal, depicting a story through its different modalities. With this motivation, we release the first large-scale memorability dataset, LAMDBA, consisting of 1749 participants and 2205 ads covering 276 brands. Running statistical tests over different participant subpopulations and ad types, we find many interesting insights into what makes an ad memorable. For e.g., we find that brands that use commercials with fast-moving scenes are more memorable than those with slower scenes (p=8e-10) and that people who use ad-blockers remember fewer ads than those who don't (p=5e-3). Next, to simulate the memorability of marketing materials for a particular audience, we present a novel model, Henry, trained to leverage real-world knowledge of LLMs and visual knowledge to predict the memorability. We test Henry on all the prominent memorability datasets in literature (both images and videos) and achieve state-of-the-art performance across all of them. Henry shows strong generalization showing better results in 0-shot on unseen datasets. Next, we propose the task of memorable ad generation and release a large-scale ad dataset, UltraLAMBDA, consisting of 4 million ads with their Henry-assigned memorability scores. We show that aligning Henry to generate memorable content improves memorability scores by more than 25%.
Authors: Xiaoqing Zhang, Jilu Zhao, Yan Li, Hao Wu, Xiangtian Zhou, Jiang Liu
Pathological myopia (PM) is the leading ocular disease for impaired vision worldwide. Clinically, the characteristic of pathology distribution in PM is global-local on the fundus image, which plays a significant role in assisting clinicians in diagnosing PM. However, most existing deep neural networks focused on designing complex architectures but rarely explored the pathology distribution prior of PM. To tackle this issue, we propose an efficient pyramid channel attention (EPCA) module, which fully leverages the potential of the clinical pathology prior of PM with pyramid pooling and multi-scale context fusion. Then, we construct EPCA-Net for automatic PM recognition based on fundus images by stacking a sequence of EPCA modules. Moreover, motivated by the recent pretraining-and-finetuning paradigm, we attempt to adapt pre-trained natural image models for PM recognition by freezing them and treating the EPCA and other attention modules as adapters. In addition, we construct a PM recognition benchmark termed PM-fundus by collecting fundus images of PM from publicly available datasets. The comprehensive experiments demonstrate the superiority of our EPCA-Net over state-of-the-art methods in the PM recognition task. The results also show that our method based on the pretraining-and-finetuning paradigm achieves competitive performance through comparisons to part of previous methods based on traditional fine-tuning paradigm with fewer tunable parameters, which has the potential to leverage more natural image foundation models to address the PM recognition task in limited medical data regime.
Authors: Wang Boya, Wang Shuo, Ye Dong, Dou Ziwen
With the frequent use of self-supervised monocular depth estimation in robotics and autonomous driving, the model's efficiency is becoming increasingly important. Most current approaches apply much larger and more complex networks to improve the precision of depth estimation. Some researchers incorporated Transformer into self-supervised monocular depth estimation to achieve better performance. However, this method leads to high parameters and high computation. We present a fully convolutional depth estimation network using contextual feature fusion. Compared to UNet++ and HRNet, we use high-resolution and low-resolution features to reserve information on small targets and fast-moving objects instead of long-range fusion. We further promote depth estimation results employing lightweight channel attention based on convolution in the decoder stage. Our method reduces the parameters without sacrificing accuracy. Experiments on the KITTI benchmark show that our method can get better results than many large models, such as Monodepth2, with only 30 parameters. The source code is available at https://github.com/boyagesmile/DNA-Depth.
Authors: Liyang Chen, Weihong Bao, Shun Lei, Boshi Tang, Zhiyong Wu, Shiyin Kang, Haozhi Huang
Speech-driven 3D facial animation aims at generating facial movements that are synchronized with the driving speech, which has been widely explored recently. Existing works mostly neglect the person-specific talking style in generation, including facial expression and head pose styles. Several works intend to capture the personalities by fine-tuning modules. However, limited training data leads to the lack of vividness. In this work, we propose AdaMesh, a novel adaptive speech-driven facial animation approach, which learns the personalized talking style from a reference video of about 10 seconds and generates vivid facial expressions and head poses. Specifically, we propose mixture-of-low-rank adaptation (MoLoRA) to fine-tune the expression adapter, which efficiently captures the facial expression style. For the personalized pose style, we propose a pose adapter by building a discrete pose prior and retrieving the appropriate style embedding with a semantic-aware pose style matrix without fine-tuning. Extensive experimental results show that our approach outperforms state-of-the-art methods, preserves the talking style in the reference video, and generates vivid facial animation. The supplementary video and code will be available at https://adamesh.github.io.
Authors: Pengchong Hu, Zhizhong Han
Learning neural implicit representations has achieved remarkable performance in 3D reconstruction from multi-view images. Current methods use volume rendering to render implicit representations into either RGB or depth images that are supervised by multi-view ground truth. However, rendering a view each time suffers from incomplete depth at holes and unawareness of occluded structures from the depth supervision, which severely affects the accuracy of geometry inference via volume rendering. To resolve this issue, we propose to learn neural implicit representations from multi-view RGBD images through volume rendering with an attentive depth fusion prior. Our prior allows neural networks to perceive coarse 3D structures from the Truncated Signed Distance Function (TSDF) fused from all depth images available for rendering. The TSDF enables accessing the missing depth at holes on one depth image and the occluded parts that are invisible from the current view. By introducing a novel attention mechanism, we allow neural networks to directly use the depth fusion prior with the inferred occupancy as the learned implicit function. Our attention mechanism works with either a one-time fused TSDF that represents a whole scene or an incrementally fused TSDF that represents a partial scene in the context of Simultaneous Localization and Mapping (SLAM). Our evaluations on widely used benchmarks including synthetic and real-world scans show our superiority over the latest neural implicit methods. Project page: https://machineperceptionlab.github.io/Attentive_DF_Prior/
Authors: Jingyi Yu, Zizhao Zhang, Shengfu Xia, Jizhang Sang
We propose a novel end-to-end pipeline for online long-range vectorized high-definition (HD) map construction using on-board camera sensors. The vectorized representation of HD maps, employing polylines and polygons to represent map elements, is widely used by downstream tasks. However, previous schemes designed with reference to dynamic object detection overlook the structural constraints within linear map elements, resulting in performance degradation in long-range scenarios. In this paper, we exploit the properties of map elements to improve the performance of map construction. We extract more accurate bird's eye view (BEV) features guided by their linear structure, and then propose a hierarchical sparse map representation to further leverage the scalability of vectorized map elements and design a progressive decoding mechanism and a supervision strategy based on this representation. Our approach, ScalableMap, demonstrates superior performance on the nuScenes dataset, especially in long-range scenarios, surpassing previous state-of-the-art model by 6.5 mAP while achieving 18.3 FPS. Code is available at https://github.com/jingy1yu/ScalableMap.
Authors: Feng Gao, Jiaxu Leng, Ji Gan, Xinbo Gao
DEtection TRansformer (DETR) and its variants (DETRs) have been successfully applied to crowded pedestrian detection, which achieved promising performance. However, we find that, in different degrees of crowded scenes, the number of DETRs' queries must be adjusted manually, otherwise, the performance would degrade to varying degrees. In this paper, we first analyze the two current query generation methods and summarize four guidelines for designing the adaptive query generation method. Then, we propose Rank-based Adaptive Query Generation (RAQG) to alleviate the problem. Specifically, we design a rank prediction head that can predict the rank of the lowest confidence positive training sample produced by the encoder. Based on the predicted rank, we design an adaptive selection method that can adaptively select coarse detection results produced by the encoder to generate queries. Moreover, to train the rank prediction head better, we propose Soft Gradient L1 Loss. The gradient of Soft Gradient L1 Loss is continuous, which can describe the relationship between the loss value and the updated value of model parameters granularly. Our method is simple and effective, which can be plugged into any DETRs to make it query-adaptive in theory. The experimental results on Crowdhuman dataset and Citypersons dataset show that our method can adaptively generate queries for DETRs and achieve competitive results. Especially, our method achieves state-of-the-art 39.4% MR on Crowdhuman dataset.
Authors: Morris Alper, Hadar Averbuch-Elor
Although the mapping between sound and meaning in human language is assumed to be largely arbitrary, research in cognitive science has shown that there are non-trivial correlations between particular sounds and meanings across languages and demographic groups, a phenomenon known as sound symbolism. Among the many dimensions of meaning, sound symbolism is particularly salient and well-demonstrated with regards to cross-modal associations between language and the visual domain. In this work, we address the question of whether sound symbolism is reflected in vision-and-language models such as CLIP and Stable Diffusion. Using zero-shot knowledge probing to investigate the inherent knowledge of these models, we find strong evidence that they do show this pattern, paralleling the well-known kiki-bouba effect in psycholinguistics. Our work provides a novel method for demonstrating sound symbolism and understanding its nature using computational tools. Our code will be made publicly available.
Authors: Kangxian Xie, Siyu Huang, Sebastian Cajas Ordone, Hanspeter Pfister, Donglai Wei
Deep-learning models have been successful in biomedical image segmentation. To generalize for real-world deployment, test-time augmentation (TTA) methods are often used to transform the test image into different versions that are hopefully closer to the training domain. Unfortunately, due to the vast diversity of instance scale and image styles, many augmented test images produce undesirable results, thus lowering the overall performance. This work proposes a new TTA framework, S$^3$-TTA, which selects the suitable image scale and style for each test image based on a transformation consistency metric. In addition, S$^3$-TTA constructs an end-to-end augmentation-segmentation joint-training pipeline to ensure a task-oriented augmentation. On public benchmarks for cell and lung segmentation, S$^3$-TTA demonstrates improvements over the prior art by 3.4% and 1.3%, respectively, by simply augmenting the input data in testing phase.
Authors: Yang Qin, Yuan Sun, Dezhong Peng, Joey Tianyi Zhou, Xi Peng, Peng Hu
Recently, image-text matching has attracted more and more attention from academia and industry, which is fundamental to understanding the latent correspondence across visual and textual modalities. However, most existing methods implicitly assume the training pairs are well-aligned while ignoring the ubiquitous annotation noise, a.k.a noisy correspondence (NC), thereby inevitably leading to a performance drop. Although some methods attempt to address such noise, they still face two challenging problems: excessive memorizing/overfitting and unreliable correction for NC, especially under high noise. To address the two problems, we propose a generalized Cross-modal Robust Complementary Learning framework (CRCL), which benefits from a novel Active Complementary Loss (ACL) and an efficient Self-refining Correspondence Correction (SCC) to improve the robustness of existing methods. Specifically, ACL exploits active and complementary learning losses to reduce the risk of providing erroneous supervision, leading to theoretically and experimentally demonstrated robustness against NC. SCC utilizes multiple self-refining processes with momentum correction to enlarge the receptive field for correcting correspondences, thereby alleviating error accumulation and achieving accurate and stable corrections. We carry out extensive experiments on three image-text benchmarks, i.e., Flickr30K, MS-COCO, and CC152K, to verify the superior robustness of our CRCL against synthetic and real-world noisy correspondences.
Authors: Xuwei Xu, Sen Wang, Yudong Chen, Yanping Zheng, Zhewei Wei, Jiajun Liu
Vision Transformers (ViTs) have revolutionized the field of computer vision, yet their deployments on resource-constrained devices remain challenging due to high computational demands. To expedite pre-trained ViTs, token pruning and token merging approaches have been developed, which aim at reducing the number of tokens involved in the computation. However, these methods still have some limitations, such as image information loss from pruned tokens and inefficiency in the token-matching process. In this paper, we introduce a novel Graph-based Token Propagation (GTP) method to resolve the challenge of balancing model efficiency and information preservation for efficient ViTs. Inspired by graph summarization algorithms, GTP meticulously propagates less significant tokens' information to spatially and semantically connected tokens that are of greater importance. Consequently, the remaining few tokens serve as a summarization of the entire token graph, allowing the method to reduce computational complexity while preserving essential information of eliminated tokens. Combined with an innovative token selection strategy, GTP can efficiently identify image tokens to be propagated. Extensive experiments have validated GTP's effectiveness, demonstrating both efficiency and performance improvements. Specifically, GTP decreases the computational complexity of both DeiT-S and DeiT-B by up to 26% with only a minimal 0.3% accuracy drop on ImageNet-1K without finetuning, and remarkably surpasses the state-of-the-art token merging method on various backbones at an even faster inference speed. The source code is available at https://github.com/Ackesnal/GTP-ViT.
Authors: Siyuan Liang, Mingli Zhu, Aishan Liu, Baoyuan Wu, Xiaochun Cao, Ee-Chien Chang
Studying backdoor attacks is valuable for model copyright protection and enhancing defenses. While existing backdoor attacks have successfully infected multimodal contrastive learning models such as CLIP, they can be easily countered by specialized backdoor defenses for MCL models. This paper reveals the threats in this practical scenario that backdoor attacks can remain effective even after defenses and introduces the \emph{\toolns} attack, which is resistant to backdoor detection and model fine-tuning defenses. To achieve this, we draw motivations from the perspective of the Bayesian rule and propose a dual-embedding guided framework for backdoor attacks. Specifically, we ensure that visual trigger patterns approximate the textual target semantics in the embedding space, making it challenging to detect the subtle parameter variations induced by backdoor learning on such natural trigger patterns. Additionally, we optimize the visual trigger patterns to align the poisoned samples with target vision features in order to hinder the backdoor unlearning through clean fine-tuning. Extensive experiments demonstrate that our attack significantly outperforms state-of-the-art baselines (+45.3% ASR) in the presence of SoTA backdoor defenses, rendering these mitigation and detection strategies virtually ineffective. Furthermore, our approach effectively attacks some more rigorous scenarios like downstream tasks. We believe that this paper raises awareness regarding the potential threats associated with the practical application of multimodal contrastive learning and encourages the development of more robust defense mechanisms.
Authors: Tingyou Li, Zixin Xu, Yong S. Chu, Xiaojing Huang, Jizhou Li
Fourier phase retrieval is essential for high-definition imaging of nanoscale structures across diverse fields, notably coherent diffraction imaging. This study presents the Single impliCit neurAl Network (SCAN), a tool built upon coordinate neural networks meticulously designed for enhanced phase retrieval performance. Remedying the drawbacks of conventional iterative methods which are easiliy trapped into local minimum solutions and sensitive to noise, SCAN adeptly connects object coordinates to their amplitude and phase within a unified network in an unsupervised manner. While many existing methods primarily use Fourier magnitude in their loss function, our approach incorporates both the predicted magnitude and phase, enhancing retrieval accuracy. Comprehensive tests validate SCAN's superiority over traditional and other deep learning models regarding accuracy and noise robustness. We also demonstrate that SCAN excels in the ptychography setting.
Authors: Gaurav Pendharkar, A.Ancy Micheal, Jason Misquitta, Ranjeesh Kaippada
Tiger conservation necessitates the strategic deployment of multifaceted initiatives encompassing the preservation of ecological habitats, anti-poaching measures, and community involvement for sustainable growth in the tiger population. With the advent of artificial intelligence, tiger surveillance can be automated using object detection. In this paper, an accurate illumination invariant framework is proposed based on EnlightenGAN and YOLOv8 for tiger detection. The fine-tuned YOLOv8 model achieves a mAP score of 61% without illumination enhancement. The illumination enhancement improves the mAP by 0.7%. The approaches elevate the state-of-the-art performance on the ATRW dataset by approximately 6% to 7%.
Authors: Aaditya Surya, Aditya Shah, Jarnell Kabore, Subash Sasikumar
This research introduces a sophisticated transfer learning model based on Google's MobileNetV2 for breast cancer tumor classification into normal, benign, and malignant categories, utilizing a dataset of 1576 ultrasound images (265 normal, 891 benign, 420 malignant). The model achieves an accuracy of 0.82, precision of 0.83, recall of 0.81, ROC-AUC of 0.94, PR-AUC of 0.88, and MCC of 0.74. It examines image intensity distributions and misclassification errors, offering improvements for future applications. Addressing dataset imbalances, the study ensures a generalizable model. This work, using a dataset from Baheya Hospital, Cairo, Egypt, compiled by Walid Al-Dhabyani et al., emphasizes MobileNetV2's potential in medical imaging, aiming to improve diagnostic precision in oncology. Additionally, the paper explores Streamlit-based deployment for real-time tumor classification, demonstrating MobileNetV2's applicability in medical imaging and setting a benchmark for future research in oncology diagnostics.
Authors: Minghui Zhang, Hanxiao Zhang, Xin You, Guang-Zhong Yang, Yun Gu
Shape modeling of volumetric data is essential for medical image analysis and computer-aided intervention. In practice, automated shape reconstruction cannot always achieve satisfactory results due to limited image resolution and a lack of sufficiently detailed shape priors used as constraints. In this paper, a unified framework is proposed for 3D shape modelling and segmentation refinement based on implicit neural networks. To learn a sharable shape prior from different instances within the same category during training, physical details of volumetric data are firstly used to construct Physical-Informed Continuous Coordinate Transform (PICCT) for implicit shape modeling. For improved shape representation, implicit shape constraints based on Signed Distance Function (SDF) are used for both instances and latent templates. For inference, a Template Interaction Module (TIM) is proposed to refine 3D shapes produced by Convolutional Neural Networks (CNNs) via deforming deep implicit templates with latent codes. Experimental results on validation datasets involving liver, pancreas and lung segmentation demonstrate the superiority of our approach in shape refinement and reconstruction. The Chamfer Distance/Earth Mover's Distance achieved by the proposed method are 0.232/0.087 for the Liver dataset, 0.128/0.069 for the Pancreas dataset, and 0.417/0.100 for the Lung Lobe dataset, respectively.
Authors: Bao Li, Zhenyu Liu, Lizhi Shao, Bensheng Qiu, Hong Bu, Jie Tian
Directly predicting human epidermal growth factor receptor 2 (HER2) status from widely available hematoxylin and eosin (HE)-stained whole slide images (WSIs) can reduce technical costs and expedite treatment selection. Accurately predicting HER2 requires large collections of multi-site WSIs. Federated learning enables collaborative training of these WSIs without gigabyte-size WSIs transportation and data privacy concerns. However, federated learning encounters challenges in addressing label imbalance in multi-site WSIs from the real world. Moreover, existing WSI classification methods cannot simultaneously exploit local context information and long-range dependencies in the site-end feature representation of federated learning. To address these issues, we present a point transformer with federated learning for multi-site HER2 status prediction from HE-stained WSIs. Our approach incorporates two novel designs. We propose a dynamic label distribution strategy and an auxiliary classifier, which helps to establish a well-initialized model and mitigate label distribution variations across sites. Additionally, we propose a farthest cosine sampling based on cosine distance. It can sample the most distinctive features and capture the long-range dependencies. Extensive experiments and analysis show that our method achieves state-of-the-art performance at four sites with a total of 2687 WSIs. Furthermore, we demonstrate that our model can generalize to two unseen sites with 229 WSIs.
Authors: Alexander Mehta, William Yang
In the task of emotion recognition from videos, a key improvement has been to focus on emotions over time rather than a single frame. There are many architectures to address this task such as GRUs, LSTMs, Self-Attention, Transformers, and Temporal Convolutional Networks (TCNs). However, these methods suffer from high memory usage, large amounts of operations, or poor gradients. We propose a method known as Neighborhood Attention with Convolutions TCN (NAC-TCN) which incorporates the benefits of attention and Temporal Convolutional Networks while ensuring that causal relationships are understood which results in a reduction in computation and memory cost. We accomplish this by introducing a causal version of Dilated Neighborhood Attention while incorporating it with convolutions. Our model achieves comparable, better, or state-of-the-art performance over TCNs, TCAN, LSTMs, and GRUs while requiring fewer parameters on standard emotion recognition datasets. We publish our code online for easy reproducibility and use in other projects.
Authors: Yi Xie, Jie Zhang, Shiqian Zhao, Tianwei Zhang, Xiaofeng Chen
While deep learning models have shown significant performance across various domains, their deployment needs extensive resources and advanced computing infrastructure. As a solution, Machine Learning as a Service (MLaaS) has emerged, lowering the barriers for users to release or productize their deep learning models. However, previous studies have highlighted potential privacy and security concerns associated with MLaaS, and one primary threat is model extraction attacks. To address this, there are many defense solutions but they suffer from unrealistic assumptions and generalization issues, making them less practical for reliable protection. Driven by these limitations, we introduce a novel defense mechanism, SAME, based on the concept of sample reconstruction. This strategy imposes minimal prerequisites on the defender's capabilities, eliminating the need for auxiliary Out-of-Distribution (OOD) datasets, user query history, white-box model access, and additional intervention during model training. It is compatible with existing active defense methods. Our extensive experiments corroborate the superior efficacy of SAME over state-of-the-art solutions. Our code is available at https://github.com/xythink/SAME.
Authors: Haeyong Kang, Jaehong Yoon, Sung Ju Hwang, Chang D. Yoo
Inspired by the Lottery Ticket Hypothesis (LTH), which highlights the existence of efficient subnetworks within larger, dense networks, a high-performing Winning Subnetwork (WSN) in terms of task performance under appropriate sparsity conditions is considered for various continual learning tasks. It leverages pre-existing weights from dense networks to achieve efficient learning in Task Incremental Learning (TIL) scenarios. In Few-Shot Class Incremental Learning (FSCIL), a variation of WSN referred to as the Soft subnetwork (SoftNet) is designed to prevent overfitting when the data samples are scarce. Furthermore, the sparse reuse of WSN weights is considered for Video Incremental Learning (VIL). The use of Fourier Subneural Operator (FSO) within WSN is considered. It enables compact encoding of videos and identifies reusable subnetworks across varying bandwidths. We have integrated FSO into different architectural frameworks for continual learning, including VIL, TIL, and FSCIL. Our comprehensive experiments demonstrate FSO's effectiveness, significantly improving task performance at various convolutional representational levels. Specifically, FSO enhances higher-layer performance in TIL and FSCIL and lower-layer performance in VIL
Authors: Kaipeng Fang, Jingkuan Song, Lianli Gao, Pengpeng Zeng, Zhi-Qi Cheng, Xiyao Li, Heng Tao Shen
The goal of Universal Cross-Domain Retrieval (UCDR) is to achieve robust performance in generalized test scenarios, wherein data may belong to strictly unknown domains and categories during training. Recently, pre-trained models with prompt tuning have shown strong generalization capabilities and attained noteworthy achievements in various downstream tasks, such as few-shot learning and video-text retrieval. However, applying them directly to UCDR may not sufficiently to handle both domain shift (i.e., adapting to unfamiliar domains) and semantic shift (i.e., transferring to unknown categories). To this end, we propose Prompting-to-Simulate (ProS), the first method to apply prompt tuning for UCDR. ProS employs a two-step process to simulate Content-aware Dynamic Prompts (CaDP) which can impact models to produce generalized features for UCDR. Concretely, in Prompt Units Learning stage, we introduce two Prompt Units to individually capture domain and semantic knowledge in a mask-and-align way. Then, in Context-aware Simulator Learning stage, we train a Content-aware Prompt Simulator under a simulated test scenarios to produce the corresponding CaDP. Extensive experiments conducted on three benchmark datasets show that our method achieves new state-of-the-art performance without bringing excessive parameters. Our method is publicly available at https://anonymous.4open.science/r/ProS
Authors: Tariq Berrada, Jakob Verbeek, Camille Couprie, Karteek Alahari
Semantic image synthesis, i.e., generating images from user-provided semantic label maps, is an important conditional image generation task as it allows to control both the content as well as the spatial layout of generated images. Although diffusion models have pushed the state of the art in generative image modeling, the iterative nature of their inference process makes them computationally demanding. Other approaches such as GANs are more efficient as they only need a single feed-forward pass for generation, but the image quality tends to suffer on large and diverse datasets. In this work, we propose a new class of GAN discriminators for semantic image synthesis that generates highly realistic images by exploiting feature backbone networks pre-trained for tasks such as image classification. We also introduce a new generator architecture with better context modeling and using cross-attention to inject noise into latent variables, leading to more diverse generated images. Our model, which we dub DP-SIMS, achieves state-of-the-art results in terms of image quality and consistency with the input label maps on ADE-20K, COCO-Stuff, and Cityscapes, surpassing recent diffusion models while requiring two orders of magnitude less compute for inference.
Authors: Riccardo Scodellaro, Ajinkya Kulkarni, Frauke Alves, Matthias Schröter
The recent successes in analyzing images with deep neural networks are almost exclusively achieved with Convolutional Neural Networks (CNNs). The training of these CNNs, and in fact of all deep neural network architectures, uses the backpropagation algorithm where the output of the network is compared with the desired result and the difference is then used to tune the weights of the network towards the desired outcome. In a 2022 preprint, Geoffrey Hinton suggested an alternative way of training which passes the desired results together with the images at the input of the network. This so called Forward Forward (FF) algorithm has up to now only been used in fully connected networks. In this paper, we show how the FF paradigm can be extended to CNNs. Our FF-trained CNN, featuring a novel spatially-extended labeling technique, achieves a classification accuracy of 99.16% on the MNIST hand-written digits dataset. We show how different hyperparameters affect the performance of the proposed algorithm and compare the results with CNN trained with the standard backpropagation approach. Furthermore, we use Class Activation Maps to investigate which type of features are learnt by the FF algorithm.
Authors: Jing Wang, Jinagyun Li, Chen Chen, Yisi Zhang, Haoran Shen, Tianxiang Zhang
The Few-Shot Segmentation (FSS) aims to accomplish the novel class segmentation task with a few annotated images. Current FSS research based on meta-learning focus on designing a complex interaction mechanism between the query and support feature. However, unlike humans who can rapidly learn new things from limited samples, the existing approach relies solely on fixed feature matching to tackle new tasks, lacking adaptability. In this paper, we propose a novel framework based on the adapter mechanism, namely Adaptive FSS, which can efficiently adapt the existing FSS model to the novel classes. In detail, we design the Prototype Adaptive Module (PAM), which utilizes accurate category information provided by the support set to derive class prototypes, enhancing class-specific information in the multi-stage representation. In addition, our approach is compatible with in diverse FSS methods with different backbones by simply inserting PAM between the layers of the encoder. Experiments demonstrate that our method effectively improves the performance of the FSS models (e.g., MSANet, HDMNet, FPTrans, and DCAMA) and achieve new state-of-the-art (SOTA) results (i.e., 72.4\% and 79.1\% mIoU on PASCAL-5$^i$ 1-shot and 5-shot settings, 52.7\% and 60.0\% mIoU on COCO-20$^i$ 1-shot and 5-shot settings). Our code can be available at https://github.com/jingw193/AdaptiveFSS.
Authors: Christian Simon, Sen He, Juan-Manuel Perez-Rua, Mengmeng Xu, Amine Benhalloum, Tao Xiang
Solving image-to-3D from a single view is an ill-posed problem, and current neural reconstruction methods addressing it through diffusion models still rely on scene-specific optimization, constraining their generalization capability. To overcome the limitations of existing approaches regarding generalization and consistency, we introduce a novel neural rendering technique. Our approach employs the signed distance function as the surface representation and incorporates generalizable priors through geometry-encoding volumes and HyperNetworks. Specifically, our method builds neural encoding volumes from generated multi-view inputs. We adjust the weights of the SDF network conditioned on an input image at test-time to allow model adaptation to novel scenes in a feed-forward manner via HyperNetworks. To mitigate artifacts derived from the synthesized views, we propose the use of a volume transformer module to improve the aggregation of image features instead of processing each viewpoint separately. Through our proposed method, dubbed as Hyper-VolTran, we avoid the bottleneck of scene-specific optimization and maintain consistency across the images generated from multiple viewpoints. Our experiments show the advantages of our proposed approach with consistent results and rapid generation.
Authors: Ingyun Lee, Wooju Lee, Hyun Myung
Deep neural networks have shown remarkable performance in image classification. However, their performance significantly deteriorates with corrupted input data. Domain generalization methods have been proposed to train robust models against out-of-distribution data. Data augmentation in the frequency domain is one of such approaches that enable a model to learn phase features to establish domain-invariant representations. This approach changes the amplitudes of the input data while preserving the phases. However, using fixed phases leads to susceptibility to phase fluctuations because amplitudes and phase fluctuations commonly occur in out-of-distribution. In this study, to address this problem, we introduce an approach using finite variation of the phases of input data rather than maintaining fixed phases. Based on the assumption that the degree of domain-invariant features varies for each phase, we propose a method to distinguish phases based on this degree. In addition, we propose a method called vital phase augmentation (VIPAug) that applies the variation to the phases differently according to the degree of domain-invariant features of given phases. The model depends more on the vital phases that contain more domain-invariant features for attaining robustness to amplitude and phase fluctuations. We present experimental evaluations of our proposed approach, which exhibited improved performance for both clean and corrupted data. VIPAug achieved SOTA performance on the benchmark CIFAR-10 and CIFAR-100 datasets, as well as near-SOTA performance on the ImageNet-100 and ImageNet datasets. Our code is available at https://github.com/excitedkid/vipaug.
Authors: Trung Tuan Dao, Duc Hong Vu, Cuong Pham, Anh Tran
The existing facial datasets, while having plentiful images at near frontal views, lack images with extreme head poses, leading to the downgraded performance of deep learning models when dealing with profile or pitched faces. This work aims to address this gap by introducing a novel dataset named Extreme Pose Face High-Quality Dataset (EFHQ), which includes a maximum of 450k high-quality images of faces at extreme poses. To produce such a massive dataset, we utilize a novel and meticulous dataset processing pipeline to curate two publicly available datasets, VFHQ and CelebV-HQ, which contain many high-resolution face videos captured in various settings. Our dataset can complement existing datasets on various facial-related tasks, such as facial synthesis with 2D/3D-aware GAN, diffusion-based text-to-image face generation, and face reenactment. Specifically, training with EFHQ helps models generalize well across diverse poses, significantly improving performance in scenarios involving extreme views, confirmed by extensive experiments. Additionally, we utilize EFHQ to define a challenging cross-view face verification benchmark, in which the performance of SOTA face recognition models drops 5-37% compared to frontal-to-frontal scenarios, aiming to stimulate studies on face recognition under severe pose conditions in the wild.
Authors: Jun Wang, Hao Ruan, Mingjie Wang, Chuanghui Zhang, Huachun Li, Jun Zhou
Over the past decade, visual gaze estimation has garnered growing attention within the research community, thanks to its wide-ranging application scenarios. While existing estimation approaches have achieved remarkable success in enhancing prediction accuracy, they primarily infer gaze directions from single-image signals and discard the huge potentials of the currently dominant text guidance. Notably, visual-language collaboration has been extensively explored across a range of visual tasks, such as image synthesis and manipulation, leveraging the remarkable transferability of large-scale Contrastive Language-Image Pre-training (CLIP) model. Nevertheless, existing gaze estimation approaches ignore the rich semantic cues conveyed by linguistic signals and priors in CLIP feature space, thereby yielding performance setbacks. In pursuit of making up this gap, we delve deeply into the text-eye collaboration protocol and introduce a novel gaze estimation framework in this paper, referred to as GazeCLIP. Specifically, we intricately design a linguistic description generator to produce text signals with coarse directional cues. Additionally, a CLIP-based backbone that excels in characterizing text-eye pairs for gaze estimation is presented. This is followed by the implementation of a fine-grained multi-modal fusion module aimed at modeling the interrelationships between heterogeneous inputs. Extensive experiments on three challenging datasets demonstrate the superiority of the proposed GazeCLIP which surpasses the previous approaches and achieves the state-of-the-art estimation accuracy.
Authors: Vardaan Pahuja, Weidi Luo, Yu Gu, Cheng-Hao Tu, Hong-You Chen, Tanya Berger-Wolf, Charles Stewart, Song Gao, Wei-Lun Chao, Yu Su
Camera traps are valuable tools in animal ecology for biodiversity monitoring and conservation. However, challenges like poor generalization to deployment at new unseen locations limit their practical application. Images are naturally associated with heterogeneous forms of context possibly in different modalities. In this work, we leverage the structured context associated with the camera trap images to improve out-of-distribution generalization for the task of species identification in camera traps. For example, a photo of a wild animal may be associated with information about where and when it was taken, as well as structured biology knowledge about the animal species. While typically overlooked by existing work, bringing back such context offers several potential benefits for better image understanding, such as addressing data scarcity and enhancing generalization. However, effectively integrating such heterogeneous context into the visual domain is a challenging problem. To address this, we propose a novel framework that reformulates species classification as link prediction in a multimodal knowledge graph (KG). This framework seamlessly integrates various forms of multimodal context for visual recognition. We apply this framework for out-of-distribution species classification on the iWildCam2020-WILDS and Snapshot Mountain Zebra datasets and achieve competitive performance with state-of-the-art approaches. Furthermore, our framework successfully incorporates biological taxonomy for improved generalization and enhances sample efficiency for recognizing under-represented species.
Authors: Yifei Chen, Chenyan Zhang, Ben Chen, Yiyu Huang, Yifei Sun, Changmiao Wang, Xianjun Fu, Yuxing Dai, Feiwei Qin, Yong Peng, Yu Gao
In standard hospital blood tests, the traditional process requires doctors to manually isolate leukocytes from microscopic images of patients' blood using microscopes. These isolated leukocytes are then categorized via automatic leukocyte classifiers to determine the proportion and volume of different types of leukocytes present in the blood samples, aiding disease diagnosis. This methodology is not only time-consuming and labor-intensive, but it also has a high propensity for errors due to factors such as image quality and environmental conditions, which could potentially lead to incorrect subsequent classifications and misdiagnosis. To address these issues, this paper proposes an innovative method of leukocyte detection: the Multi-level Feature Fusion and Deformable Self-attention DETR (MFDS-DETR). To tackle the issue of leukocyte scale disparity, we designed the High-level Screening-feature Fusion Pyramid (HS-FPN), enabling multi-level fusion. This model uses high-level features as weights to filter low-level feature information via a channel attention module and then merges the screened information with the high-level features, thus enhancing the model's feature expression capability. Further, we address the issue of leukocyte feature scarcity by incorporating a multi-scale deformable self-attention module in the encoder and using the self-attention and cross-deformable attention mechanisms in the decoder, which aids in the extraction of the global features of the leukocyte feature maps. The effectiveness, superiority, and generalizability of the proposed MFDS-DETR method are confirmed through comparisons with other cutting-edge leukocyte detection models using the private WBCDD, public LISC and BCCD datasets. Our source code and private WBCCD dataset are available at https://github.com/JustlfC03/MFDS-DETR.
Authors: Qiuhui Chen, Yi Hong
Medical data collected for making a diagnostic decision are typically multi-modal and provide complementary perspectives of a subject. A computer-aided diagnosis system welcomes multi-modal inputs; however, how to effectively fuse such multi-modal data is a challenging task and attracts a lot of attention in the medical research field. In this paper, we propose a transformer-based framework, called Alifuse, for aligning and fusing multi-modal medical data. Specifically, we convert images and unstructured and structured texts into vision and language tokens, and use intramodal and intermodal attention mechanisms to learn holistic representations of all imaging and non-imaging data for classification. We apply Alifuse to classify Alzheimer's disease and obtain state-of-the-art performance on five public datasets, by outperforming eight baselines. The source code will be available online later.
Authors: Ningyu Zhang, Yunzhi Yao, Bozhong Tian, Peng Wang, Shumin Deng, Mengru Wang, Zekun Xi, Shengyu Mao, Jintian Zhang, Yuansheng Ni, Siyuan Cheng, Ziwen Xu, Xin Xu, Jia-Chen Gu, Yong Jiang, Pengjun Xie, Fei Huang, Lei Liang, Zhiqiang Zhang, Xiaowei Zhu, Jun Zhou, Huajun Chen
Large Language Models (LLMs) have shown extraordinary capabilities in understanding and generating text that closely mirrors human communication. However, a primary limitation lies in the significant computational demands during training, arising from their extensive parameterization. This challenge is further intensified by the dynamic nature of the world, necessitating frequent updates to LLMs to correct outdated information or integrate new knowledge, thereby ensuring their continued relevance. Note that many applications demand continual model adjustments post-training to address deficiencies or undesirable behaviors. There is an increasing interest in efficient, lightweight methods for on-the-fly model modifications. To this end, recent years have seen a burgeoning in the techniques of knowledge editing for LLMs, which aim to efficiently modify LLMs' behaviors within specific domains while preserving overall performance across various inputs. In this paper, we first define the knowledge editing problem and then provide a comprehensive review of cutting-edge approaches. Drawing inspiration from educational and cognitive research theories, we propose a unified categorization criterion that classifies knowledge editing methods into three groups: resorting to external knowledge, merging knowledge into the model, and editing intrinsic knowledge. Furthermore, we introduce a new benchmark, KnowEdit, for a comprehensive empirical evaluation of representative knowledge editing approaches. Additionally, we provide an in-depth analysis of knowledge location, which can provide a deeper understanding of the knowledge structures inherent within LLMs. Finally, we discuss several potential applications of knowledge editing, outlining its broad and impactful implications.
Authors: Michalis Pistos, Gang Li, Weili Lin, Dinggang Shen, Islem Rekik
The understanding of the convoluted evolution of infant brain networks during the first postnatal year is pivotal for identifying the dynamics of early brain connectivity development. Existing deep learning solutions suffer from three major limitations. First, they cannot generalize to multi-trajectory prediction tasks, where each graph trajectory corresponds to a particular imaging modality or connectivity type (e.g., T1-w MRI). Second, existing models require extensive training datasets to achieve satisfactory performance which are often challenging to obtain. Third, they do not efficiently utilize incomplete time series data. To address these limitations, we introduce FedGmTE-Net++, a federated graph-based multi-trajectory evolution network. Using the power of federation, we aggregate local learnings among diverse hospitals with limited datasets. As a result, we enhance the performance of each hospital's local generative model, while preserving data privacy. The three key innovations of FedGmTE-Net++ are: (i) presenting the first federated learning framework specifically designed for brain multi-trajectory evolution prediction in a data-scarce environment, (ii) incorporating an auxiliary regularizer in the local objective function to exploit all the longitudinal brain connectivity within the evolution trajectory and maximize data utilization, (iii) introducing a two-step imputation process, comprising a preliminary KNN-based precompletion followed by an imputation refinement step that employs regressors to improve similarity scores and refine imputations. Our comprehensive experimental results showed the outperformance of FedGmTE-Net++ in brain multi-trajectory prediction from a single baseline graph in comparison with benchmark methods.
Authors: Wentao Zhu
Vision transformers (ViTs) have achieved promising results on a variety of Computer Vision tasks, however their quadratic complexity in the number of input tokens has limited their application specially in resource-constrained settings. Previous approaches that employ gradual token reduction to address this challenge assume that token redundancy in one layer implies redundancy in all the following layers. We empirically demonstrate that this assumption is often not correct, i.e., tokens that are redundant in one layer can be useful in later layers. We employ this key insight to propose a novel token propagation controller (TPC) that incorporates two different token-distributions, i.e., pause probability and restart probability to control the reduction and reuse of tokens respectively, which results in more efficient token utilization. To improve the estimates of token distributions, we propose a smoothing mechanism that acts as a regularizer and helps remove noisy outliers. Furthermore, to improve the training-stability of our proposed TPC, we introduce a model stabilizer that is able to implicitly encode local image structures and minimize accuracy fluctuations during model training. We present extensive experimental results on the ImageNet-1K dataset using DeiT, LV-ViT and Swin models to demonstrate the effectiveness of our proposed method. For example, compared to baseline models, our proposed method improves the inference speed of the DeiT-S by 250% while increasing the classification accuracy by 1.0%.
Authors: Haopeng Li, Andong Deng, Qiuhong Ke, Jun Liu, Hossein Rahmani, Yulan Guo, Bernt Schiele, Chen Chen
Reasoning over sports videos for question answering is an important task with numerous applications, such as player training and information retrieval. However, this task has not been explored due to the lack of relevant datasets and the challenging nature it presents. Most datasets for video question answering (VideoQA) focus mainly on general and coarse-grained understanding of daily-life videos, which is not applicable to sports scenarios requiring professional action understanding and fine-grained motion analysis. In this paper, we introduce the first dataset, named Sports-QA, specifically designed for the sports VideoQA task. The Sports-QA dataset includes various types of questions, such as descriptions, chronologies, causalities, and counterfactual conditions, covering multiple sports. Furthermore, to address the characteristics of the sports VideoQA task, we propose a new Auto-Focus Transformer (AFT) capable of automatically focusing on particular scales of temporal information for question answering. We conduct extensive experiments on Sports-QA, including baseline studies and the evaluation of different methods. The results demonstrate that our AFT achieves state-of-the-art performance.
Authors: Fanda Fan, Chunjie Luo, Wanling Gao, Jianfeng Zhan
The burgeoning field of Artificial Intelligence Generated Content (AIGC) is witnessing rapid advancements, particularly in video generation. This paper introduces AIGCBench, a pioneering comprehensive and scalable benchmark designed to evaluate a variety of video generation tasks, with a primary focus on Image-to-Video (I2V) generation. AIGCBench tackles the limitations of existing benchmarks, which suffer from a lack of diverse datasets, by including a varied and open-domain image-text dataset that evaluates different state-of-the-art algorithms under equivalent conditions. We employ a novel text combiner and GPT-4 to create rich text prompts, which are then used to generate images via advanced Text-to-Image models. To establish a unified evaluation framework for video generation tasks, our benchmark includes 11 metrics spanning four dimensions to assess algorithm performance. These dimensions are control-video alignment, motion effects, temporal consistency, and video quality. These metrics are both reference video-dependent and video-free, ensuring a comprehensive evaluation strategy. The evaluation standard proposed correlates well with human judgment, providing insights into the strengths and weaknesses of current I2V algorithms. The findings from our extensive experiments aim to stimulate further research and development in the I2V field. AIGCBench represents a significant step toward creating standardized benchmarks for the broader AIGC landscape, proposing an adaptable and equitable framework for future assessments of video generation tasks. We have open-sourced the dataset and evaluation code on the project website: https://www.benchcouncil.org/AIGCBench.
Authors: Haowen Zheng, Dong Cao, Jintao Xu, Rui Ai, Weihao Gu, Yang Yang, Yanyan Liang
Striking a balance between precision and efficiency presents a prominent challenge in the bird's-eye-view (BEV) 3D object detection. Although previous camera-based BEV methods achieved remarkable performance by incorporating long-term temporal information, most of them still face the problem of low efficiency. One potential solution is knowledge distillation. Existing distillation methods only focus on reconstructing spatial features, while overlooking temporal knowledge. To this end, we propose TempDistiller, a Temporal knowledge Distiller, to acquire long-term memory from a teacher detector when provided with a limited number of frames. Specifically, a reconstruction target is formulated by integrating long-term temporal knowledge through self-attention operation applied to feature teachers. Subsequently, novel features are generated for masked student features via a generator. Ultimately, we utilize this reconstruction target to reconstruct the student features. In addition, we also explore temporal relational knowledge when inputting full frames for the student model. We verify the effectiveness of the proposed method on the nuScenes benchmark. The experimental results show our method obtain an enhancement of +1.6 mAP and +1.1 NDS compared to the baseline, a speed improvement of approximately 6 FPS after compressing temporal knowledge, and the most accurate velocity estimation.
Authors: Xuehao Gao, Yang Yang, Zhenyu Xie, Shaoyi Du, Zhongqian Sun, Yang Wu
In this paper, we propose a novel cascaded diffusion-based generative framework for text-driven human motion synthesis, which exploits a strategy named GradUally Enriching SyntheSis (GUESS as its abbreviation). The strategy sets up generation objectives by grouping body joints of detailed skeletons in close semantic proximity together and then replacing each of such joint group with a single body-part node. Such an operation recursively abstracts a human pose to coarser and coarser skeletons at multiple granularity levels. With gradually increasing the abstraction level, human motion becomes more and more concise and stable, significantly benefiting the cross-modal motion synthesis task. The whole text-driven human motion synthesis problem is then divided into multiple abstraction levels and solved with a multi-stage generation framework with a cascaded latent diffusion model: an initial generator first generates the coarsest human motion guess from a given text description; then, a series of successive generators gradually enrich the motion details based on the textual description and the previous synthesized results. Notably, we further integrate GUESS with the proposed dynamic multi-condition fusion mechanism to dynamically balance the cooperative effects of the given textual condition and synthesized coarse motion prompt in different generation stages. Extensive experiments on large-scale datasets verify that GUESS outperforms existing state-of-the-art methods by large margins in terms of accuracy, realisticness, and diversity. Code is available at https://github.com/Xuehao-Gao/GUESS.
Authors: Yiran Song, Qianyu Zhou, Xiangtai Li, Deng-Ping Fan, Xuequan Lu, Lizhuang Ma
In this paper, we address the challenge of image resolution variation for the Segment Anything Model (SAM). SAM, known for its zero-shot generalizability, exhibits a performance degradation when faced with datasets with varying image sizes. Previous approaches tend to resize the image to a fixed size or adopt structure modifications, hindering the preservation of SAM's rich prior knowledge. Besides, such task-specific tuning necessitates a complete retraining of the model, which is cost-expensive and unacceptable for deployment in the downstream tasks. In this paper, we reformulate this issue as a length extrapolation problem, where token sequence length varies while maintaining a consistent patch size for images of different sizes. To this end, we propose Scalable Bias-Mode Attention Mask (BA-SAM) to enhance SAM's adaptability to varying image resolutions while eliminating the need for structure modifications. Firstly, we introduce a new scaling factor to ensure consistent magnitude in the attention layer's dot product values when the token sequence length changes. Secondly, we present a bias-mode attention mask that allows each token to prioritize neighboring information, mitigating the impact of untrained distant information. Our BA-SAM demonstrates efficacy in two scenarios: zero-shot and fine-tuning. Extensive evaluation on diverse datasets, including DIS5K, DUTS, ISIC, COD10K, and COCO, reveals its ability to significantly mitigate performance degradation in the zero-shot setting and achieve state-of-the-art performance with minimal fine-tuning. Furthermore, we propose a generalized model and benchmark, showcasing BA-SAM's generalizability across all four datasets simultaneously.
Authors: Jai Prakash Veerla, Poojitha Thota, Partha Sai Guttikonda, Shirin Nilizadeh, Jacob M. Luber
In the dynamic landscape of medical artificial intelligence, this study explores the vulnerabilities of the Pathology Language-Image Pretraining (PLIP) model, a Vision Language Foundation model, under targeted adversarial conditions. Leveraging the Kather Colon dataset with 7,180 H&E images across nine tissue types, our investigation employs Projected Gradient Descent (PGD) adversarial attacks to intentionally induce misclassifications. The outcomes reveal a 100% success rate in manipulating PLIP's predictions, underscoring its susceptibility to adversarial perturbations. The qualitative analysis of adversarial examples delves into the interpretability challenges, shedding light on nuanced changes in predictions induced by adversarial manipulations. These findings contribute crucial insights into the interpretability, domain adaptation, and trustworthiness of Vision Language Models in medical imaging. The study emphasizes the pressing need for robust defenses to ensure the reliability of AI models.
Authors: Yuta Okuyama, Yuki Endo, Yoshihiro Kanamori
Pose and body shape editing in a human image has received increasing attention. However, current methods often struggle with dataset biases and deteriorate realism and the person's identity when users make large edits. We propose a one-shot approach that enables large edits with identity preservation. To enable large edits, we fit a 3D body model, project the input image onto the 3D model, and change the body's pose and shape. Because this initial textured body model has artifacts due to occlusion and the inaccurate body shape, the rendered image undergoes a diffusion-based refinement, in which strong noise destroys body structure and identity whereas insufficient noise does not help. We thus propose an iterative refinement with weak noise, applied first for the whole body and then for the face. We further enhance the realism by fine-tuning text embeddings via self-supervised learning. Our quantitative and qualitative evaluations demonstrate that our method outperforms other existing methods across various datasets.
Authors: Yuxin Yang, Pengfei Zhu, Mengshi Qi, Huadong Ma
Human trajectory forecasting is a critical challenge in fields such as robotics and autonomous driving. Due to the inherent uncertainty of human actions and intentions in real-world scenarios, various unexpected occurrences may arise. To uncover latent motion patterns in human behavior, we introduce a novel memory-based method, named Motion Pattern Priors Memory Network. Our method involves constructing a memory bank derived from clustered prior knowledge of motion patterns observed in the training set trajectories. We introduce an addressing mechanism to retrieve the matched pattern and the potential target distributions for each prediction from the memory bank, which enables the identification and retrieval of natural motion patterns exhibited by agents, subsequently using the target priors memory token to guide the diffusion model to generate predictions. Extensive experiments validate the effectiveness of our approach, achieving state-of-the-art trajectory prediction accuracy. The code will be made publicly available.