As the number and complexity of sensors in autonomous vehicles continue to rise,multimodal fusionbased object detection algorithms are increasingly being used to detect 3D environmental information,significantly advan...As the number and complexity of sensors in autonomous vehicles continue to rise,multimodal fusionbased object detection algorithms are increasingly being used to detect 3D environmental information,significantly advancing the development of perception technology in autonomous driving.To further promote the development of fusion algorithms and improve detection performance,this paper discusses the advantages and recent advancements of multimodal fusion-based object detection algorithms.Starting fromsingle-modal sensor detection,the paper provides a detailed overview of typical sensors used in autonomous driving and introduces object detection methods based on images and point clouds.For image-based detection methods,they are categorized into monocular detection and binocular detection based on different input types.For point cloud-based detection methods,they are classified into projection-based,voxel-based,point cluster-based,pillar-based,and graph structure-based approaches based on the technical pathways for processing point cloud features.Additionally,multimodal fusion algorithms are divided into Camera-LiDAR fusion,Camera-Radar fusion,Camera-LiDAR-Radar fusion,and other sensor fusion methods based on the types of sensors involved.Furthermore,the paper identifies five key future research directions in this field,aiming to provide insights for researchers engaged in multimodal fusion-based object detection algorithms and to encourage broader attention to the research and application of multimodal fusion-based object detection.展开更多
Visual object tracking(VOT),aiming to track a target object in a continuous video,is a fundamental and critical task in computer vision.However,the reliance on third-party resources(e.g.,dataset)for training poses con...Visual object tracking(VOT),aiming to track a target object in a continuous video,is a fundamental and critical task in computer vision.However,the reliance on third-party resources(e.g.,dataset)for training poses concealed threats to the security of VOT models.In this paper,we reveal that VOT models are vulnerable to a poison-only and targeted backdoor attack,where the adversary can achieve arbitrary tracking predictions by manipulating only part of the training data.Specifically,we first define and formulate three different variants of the targeted attacks:size-manipulation,trajectory-manipulation,and hybrid attacks.To implement these,we introduce Random Video Poisoning(RVP),a novel poison-only strategy that exploits temporal correlations within video data by poisoning entire video sequences.Extensive experiments demonstrate that RVP effectively injects controllable backdoors,enabling precise manipulation of tracking behavior upon trigger activation,while maintaining high performance on benign data,thus ensuring stealth.Our findings not only expose significant vulnerabilities but also highlight that the underlying principles could be adapted for beneficial uses,such as dataset watermarking for copyright protection.展开更多
To improve small object detection and trajectory estimation from an aerial moving perspective,we propose the Aerial View Attention-PRB(AVA-PRB)model.AVA-PRB integrates two attention mechanisms—Coordinate Attention(CA...To improve small object detection and trajectory estimation from an aerial moving perspective,we propose the Aerial View Attention-PRB(AVA-PRB)model.AVA-PRB integrates two attention mechanisms—Coordinate Attention(CA)and the Convolutional Block Attention Module(CBAM)—to enhance detection accuracy.Additionally,Shape-IoU is employed as the loss function to refine localization precision.Our model further incorporates an adaptive feature fusion mechanism,which optimizes multi-scale object representation,ensuring robust tracking in complex aerial environments.We evaluate the performance of AVA-PRB on two benchmark datasets:Aerial Person Detection and VisDrone2019-Det.The model achieves 60.9%mAP@0.5 on the Aerial Person Detection dataset,and 51.2%mAP@0.5 on VisDrone2019-Det,demonstrating its effectiveness in aerial object detection.Beyond detection,we propose a novel trajectory estimation method that improves movement path prediction under aerial motion.Experimental results indicate that our approach reduces path deviation by up to 64%,effectively mitigating errors caused by rapid camera movements and background variations.By optimizing feature extraction and enhancing spatialtemporal coherence,our method significantly improves object tracking under aerial moving perspectives.This research addresses the limitations of fixed-camera tracking,enhancing flexibility and accuracy in aerial tracking applications.The proposed approach has broad potential for real-world applications,including surveillance,traffic monitoring,and environmental observation.展开更多
Advancements in animal behavior quantification methods have driven the development of computational ethology,enabling fully automated behavior analysis.Existing multianimal pose estimation workflows rely on tracking-b...Advancements in animal behavior quantification methods have driven the development of computational ethology,enabling fully automated behavior analysis.Existing multianimal pose estimation workflows rely on tracking-bydetection frameworks for either bottom-up or top-down approaches,requiring retraining to accommodate diverse animal appearances.This study introduces InteBOMB,an integrated workflow that enhances top-down approaches by incorporating generic object tracking,eliminating the need for prior knowledge of target animals while maintaining broad generalizability.InteBOMB includes two key strategies for tracking and segmentation in laboratory environments and two techniques for pose estimation in natural settings.The“background enhancement”strategy optimizesforeground-backgroundcontrastiveloss,generating more discriminative correlation maps.The“online proofreading”strategy stores human-in-the-loop long-term memory and dynamic short-term memory,enabling adaptive updates to object visual features.The“automated labeling suggestion”technique reuses the visual features saved during tracking to identify representative frames for training set labeling.Additionally,the“joint behavior analysis”technique integrates these features with multimodal data,expanding the latent space for behavior classification and clustering.To evaluate the framework,six datasets of mice and six datasets of nonhuman primates were compiled,covering laboratory and natural scenes.Benchmarking results demonstrated a24%improvement in zero-shot generic tracking and a 21%enhancement in joint latent space performance across datasets,highlighting the effectiveness of this approach in robust,generalizable behavior analysis.展开更多
Visual object tracking plays a crucial role in computer vision.In recent years,researchers have proposed various methods to achieve high-performance object tracking.Among these,methods based on Transformers have becom...Visual object tracking plays a crucial role in computer vision.In recent years,researchers have proposed various methods to achieve high-performance object tracking.Among these,methods based on Transformers have become a research hotspot due to their ability to globally model and contextualize information.However,current Transformer-based object tracking methods still face challenges such as low tracking accuracy and the presence of redundant feature information.In this paper,we introduce self-calibration multi-head self-attention Transformer(SMSTracker)as a solution to these challenges.It employs a hybrid tensor decomposition self-organizing multihead self-attention transformermechanism,which not only compresses and accelerates Transformer operations but also significantly reduces redundant data,thereby enhancing the accuracy and efficiency of tracking.Additionally,we introduce a self-calibration attention fusion block to resolve common issues of attention ambiguities and inconsistencies found in traditional trackingmethods,ensuring the stability and reliability of tracking performance across various scenarios.By integrating a hybrid tensor decomposition approach with a self-organizingmulti-head self-attentive transformer mechanism,SMSTracker enhances the efficiency and accuracy of the tracking process.Experimental results show that SMSTracker achieves competitive performance in visual object tracking,promising more robust and efficient tracking systems,demonstrating its potential to providemore robust and efficient tracking solutions in real-world applications.展开更多
Significant advancements have beenwitnessed in visual tracking applications leveragingViT in recent years,mainly due to the formidablemodeling capabilities of Vision Transformer(ViT).However,the strong performance of ...Significant advancements have beenwitnessed in visual tracking applications leveragingViT in recent years,mainly due to the formidablemodeling capabilities of Vision Transformer(ViT).However,the strong performance of such trackers heavily relies on ViT models pretrained for long periods,limitingmore flexible model designs for tracking tasks.To address this issue,we propose an efficient unsupervised ViT pretraining method for the tracking task based on masked autoencoders,called TrackMAE.During pretraining,we employ two shared-parameter ViTs,serving as the appearance encoder and motion encoder,respectively.The appearance encoder encodes randomly masked image data,while the motion encoder encodes randomly masked pairs of video frames.Subsequently,an appearance decoder and a motion decoder separately reconstruct the original image data and video frame data at the pixel level.In this way,ViT learns to understand both the appearance of images and the motion between video frames simultaneously.Experimental results demonstrate that ViT-Base and ViT-Large models,pretrained with TrackMAE and combined with a simple tracking head,achieve state-of-the-art(SOTA)performance without additional design.Moreover,compared to the currently popular MAE pretraining methods,TrackMAE consumes only 1/5 of the training time,which will facilitate the customization of diverse models for tracking.For instance,we additionally customize a lightweight ViT-XS,which achieves SOTA efficient tracking performance.展开更多
In order to solve the problem of small object size and low detection accuracy under the unmanned aerial vehicle(UAV)platform,the object detection algorithm based on deep aggregation network and high-resolution fusion ...In order to solve the problem of small object size and low detection accuracy under the unmanned aerial vehicle(UAV)platform,the object detection algorithm based on deep aggregation network and high-resolution fusion module is studied.Furthermore,a joint network of object detection and feature extraction is studied to construct a real-time multi-object tracking algorithm.For the problem of object association failure caused by UAV movement,image registration is applied to multi-object tracking and a camera motion discrimination model is proposed to improve the speed of the multi-object tracking algorithm.The simulation results show that the algorithm proposed in this study can improve the accuracy of multi-object tracking under the UAV platform,and effectively solve the problem of association failure caused by UAV movement.展开更多
Crack length measurement algorithms based on computer vision have shown promising engineering application prospects in the field of aircraft fatigue crack monitoring.However,due to the complexity of the monitoring env...Crack length measurement algorithms based on computer vision have shown promising engineering application prospects in the field of aircraft fatigue crack monitoring.However,due to the complexity of the monitoring environment,the subtle visual features of small fatigue cracks,and the impact of structural elastic deformation,directly applying object segmentation algorithms often results in significant measurement errors.Therefore,this paper proposes a high-precision crack length measurement method based on Bidirectional Target Tracking Model(Bi2TM),which integrates crack tip localization,interference identification,and length compensation.First,a general object segmentation model is used to perform rough crack segmentation.Then,the Bi2TM network,combined with the visual features of the structure in different stress states,is employed to track the bidirectional position of the crack tip in the“open”and“closed”states.This ultimately enables interference identification within the rough segmented crack region,achieving highprecision length measurement.In a high-interference environment of aircraft fatigue testing,the proposed method is used to measure 1000 crack images ranging from 1 mm to 11 mm.For more than 90%of the samples,the measurement error is less than 5 pixels,demonstrating significant advantages over the existing methods.展开更多
Multi-modal 3D object detection has achieved remarkable progress,but it is often limited in practical industrial production because of its high cost and low efficiency.The multi-view camera-based method provides a fea...Multi-modal 3D object detection has achieved remarkable progress,but it is often limited in practical industrial production because of its high cost and low efficiency.The multi-view camera-based method provides a feasible solution due to its low cost.However,camera data lacks geometric depth,and only using camera data to obtain high accuracy is challenging.This paper proposes a multi-modal Bird-Eye-View(BEV)distillation framework(MMDistill)to make a trade-off between them.MMDistill is a carefully crafted two-stage distillation framework based on teacher and student models for learning cross-modal knowledge and generating multi-modal features.It can improve the performance of unimodal detectors without introducing additional costs during inference.Specifically,our method can effectively solve the cross-gap caused by the heterogeneity between data.Furthermore,we further propose a Light Detection and Ranging(LiDAR)-guided geometric compensation module,which can assist the student model in obtaining effective geometric features and reduce the gap between different modalities.Our proposed method generally requires fewer computational resources and faster inference speed than traditional multi-modal models.This advancement enables multi-modal technology to be applied more widely in practical scenarios.Through experiments,we validate the effectiveness and superiority of MMDistill on the nuScenes dataset,achieving an improvement of 4.1%mean Average Precision(mAP)and 4.6%NuScenes Detection Score(NDS)over the baseline detector.In addition,we also present detailed ablation studies to validate our method.展开更多
Visual object tracking is an important issue that has received long-term attention in computer vision.The ability to effectively handle occlusion,especially severe occlusion,is an important aspect of evaluating the pe...Visual object tracking is an important issue that has received long-term attention in computer vision.The ability to effectively handle occlusion,especially severe occlusion,is an important aspect of evaluating the performance of object tracking algorithms in long-term tracking,and is of great significance to improving the robustness of object tracking algorithms.However,most object tracking algorithms lack a processing mechanism specifically for occlusion.In the case of occlusion,due to the lack of target information,it is necessary to predict the target position based on the motion trajectory.Kalman filtering and particle filtering can effectively predict the target motion state based on the historical motion information.A single object tracking method,called probabilistic discriminative model prediction(PrDiMP),is based on the spatial attention mechanism in complex scenes and occlusions.In order to improve the performance of PrDiMP,Kalman filtering,particle filtering and linear filtering are introduced.First,for the occlusion situation,Kalman filtering and particle filtering are respectively introduced to predict the object position,thereby replacing the detection result of the original tracking algorithm and stopping recursion of target model.Second,for detection-jump problem of similar objects in complex scenes,a linear filtering window is added.The evaluation results on the three datasets,including GOT-10k,UAV123 and LaSOT,and the visualization results on several videos,show that our algorithms have improved tracking performance under occlusion and the detection-jump is effectively suppressed.展开更多
Visual object-tracking is a fundamental task applied in many applications of computer vision. Particle filter is one of the techniques which has been widely used in object tracking. Due to the virtue of extendability ...Visual object-tracking is a fundamental task applied in many applications of computer vision. Particle filter is one of the techniques which has been widely used in object tracking. Due to the virtue of extendability and flexibility on both linear and non-linear environments, various particle filter-based trackers have been proposed in the literature. However, the conventional approach cannot handle very large videos efficiently in the current data intensive information age. In this work, a parallelized particle filter is provided in a distributed framework provided by the Hadoop/Map-Reduce infrastructure to tackle object-tracking tasks. The experiments indicate that the proposed algorithm has a better convergence and accuracy as compared to the traditional particle filter. The computational power and the scalability of the proposed particle filter in single object tracking have been enhanced as well.展开更多
Droplet microfluidics,which encapsulates individual cells within separate microreactors,has become an essential tool for single-cell phenotypic and genotypic analysis.However,the efficiency of single-cell encapsulatio...Droplet microfluidics,which encapsulates individual cells within separate microreactors,has become an essential tool for single-cell phenotypic and genotypic analysis.However,the efficiency of single-cell encapsulation is limited by the Poisson distribution governing the encapsulation process,resulting in most droplets being either empty or containing multiple cells.Traditional single-cell sorting methods typically rely on fluorescence labeling for identification,but this approach not only increases experimental costs and complexity but can also impact cell viability.Additionally,current label-free sorting methods still encounter difficulties in accurately detecting multicellular droplets and small cellular aggregates.To address these challenges,this paper proposes an intelligent sorting system that combines YOLOv8 object detection and BoTSORT tracking algorithms.This system enables real-time analysis of droplet images,facilitating precise identification,counting,and automated sorting of target droplets.To validate the system’s performance,polystyrene microspheres were used to simulate real cells in sorting tests.The results demonstrated that,under label-free conditions,the system significantly outperformed traditional fluorescence labeling methods in both classification accuracy and sorting efficiency.This system provides an effective,label-free solution for cell sorting,with potential applications in precision medicine,single-cell sequencing,and drug screening.展开更多
An improved estimation of motion vectors of feature points is proposed for tracking moving objects of dynamic image sequence. Feature points are firstly extracted by the improved minimum intensity change （MIC） algor...An improved estimation of motion vectors of feature points is proposed for tracking moving objects of dynamic image sequence. Feature points are firstly extracted by the improved minimum intensity change （MIC） algorithm. The matching points of these feature points are then determined by adaptive rood pattern searching. Based on the random sample consensus （RANSAC） method, the background motion is finally compensated by the parameters of an affine transform of the background motion. With reasonable morphological filtering, the moving objects are completely extracted from the background, and then tracked accurately. Experimental results show that the improved method is successful on the motion background compensation and offers great promise in tracking moving objects of the dynamic image sequence.展开更多
This paper describes a new framework for object detection and tracking of AUV including underwater acoustic data interpolation, underwater acoustic images segmentation and underwater objects tracking. This framework i...This paper describes a new framework for object detection and tracking of AUV including underwater acoustic data interpolation, underwater acoustic images segmentation and underwater objects tracking. This framework is applied to the design of vision-based method for AUV based on the forward looking sonar sensor. First, the real-time data flow （underwater acoustic images） is pre-processed to form the whole underwater acoustic image, and the relevant position information of objects is extracted and determined. An improved method of double threshold segmentation is proposed to resolve the problem that the threshold cannot be adjusted adaptively in the traditional method. Second, a representation of region information is created in light of the Gaussian particle filter. The weighted integration strategy combining the area and invariant moment is proposed to perfect the weight of particles and to enhance the tracking robustness. Results obtained on the real acoustic vision platform of AUV during sea trials are displayed and discussed. They show that the proposed method can detect and track the moving objects underwater online, and it is effective and robust.展开更多
This paper addresses the problem of real-time object tracking for unmanned aerial vehicles. We consider the task of object tracking as a classification problem. Training a good classifier always needs a huge number of...This paper addresses the problem of real-time object tracking for unmanned aerial vehicles. We consider the task of object tracking as a classification problem. Training a good classifier always needs a huge number of samples, which is always time-consuming and not suitable for realtime applications. In this paper, we transform the large-scale least-squares problem in the spatial domain to a series of small-scale least-squares problems with constraints in the Fourier domain using the correlation filter technique. Then, this problem is efficiently solved by two stages. In the first stage, a fast method based on recursive least squares is used to solve the correlation filter problem without constraints in the Fourier domain. In the second stage, a weight matrix is constructed to prune the solution attained in the first stage to approach the constraints in the spatial domain. Then, the pruned classifier is used for tracking. To evaluate proposed tracker’s performance, comprehensive experiments are conducted on challenging aerial sequences in the UAV123 dataset. Experimental results demonstrate that proposed approach achieves a state-ofthe-art tracking performance in aerial sequences and operates at a mean speed of beyond 40 frames/s. For further analysis of proposed tracker’s robustness, extensive experiments are also performed on recent benchmarks OTB50, OTB100, and VOT2016.展开更多
A method for moving object recognition and tracking in the intelligent traffic monitoring system is presented. For the shortcomings and deficiencies of the frame-subtraction method, a redundant discrete wavelet transf...A method for moving object recognition and tracking in the intelligent traffic monitoring system is presented. For the shortcomings and deficiencies of the frame-subtraction method, a redundant discrete wavelet transform （RDWT） based moving object recognition algorithm is put forward, which directly detects moving objects in the redundant discrete wavelet transform domain. An improved adaptive mean-shift algorithm is used to track the moving object in the follow up frames. Experimental results show that the algorithm can effectively extract the moving object, even though the object is similar to the background, and the results are better than the traditional frame-subtraction method. The object tracking is accurate without the impact of changes in the size of the object. Therefore the algorithm has a certain practical value and prospect.展开更多
为了解决现有的目标跟踪算法在煤矿复杂环境下存在精度低和实时性差的问题,基于Tracking by Detection(TBD)范式,提出了YOLO-FasterNet+ByteTrack的煤矿人员跟踪算法。首先,构建FasterNet-Block特征提取模块改进YOLOv7的Backbone,提升...为了解决现有的目标跟踪算法在煤矿复杂环境下存在精度低和实时性差的问题,基于Tracking by Detection(TBD)范式,提出了YOLO-FasterNet+ByteTrack的煤矿人员跟踪算法。首先,构建FasterNet-Block特征提取模块改进YOLOv7的Backbone,提升目标检测阶段的实时性;然后,通过在Neck中引入CBAM注意力机制,提升模型在复杂场景下的特征感知能力;接着,在目标检测的解码阶段引入Soft-NMS,优化模型在人员交叠场景下的检测精度;最后,在目标跟踪阶段,针对人员重叠和遮挡导致的目标ID翻转问题,设计了一种融合GRU和卡尔曼滤波的多目标运动特征预测机制,有效提升了煤矿人员跟踪的准确性。实验结果表明:YOLOFasterNet在煤矿人员数据集上相对于YOLOv7的平均精度提高了3.6%,检测速度提升了8.2FPS;在自定义跟踪数据集GBMOT上,所提目标跟踪算法相对于ByteTrack,MOTA值提升了1.7%,IDSW减少了149次。展开更多
Collaborative Robotics is one of the high-interest research topics in the area of academia and industry.It has been progressively utilized in numerous applications,particularly in intelligent surveillance systems.It a...Collaborative Robotics is one of the high-interest research topics in the area of academia and industry.It has been progressively utilized in numerous applications,particularly in intelligent surveillance systems.It allows the deployment of smart cameras or optical sensors with computer vision techniques,which may serve in several object detection and tracking tasks.These tasks have been considered challenging and high-level perceptual problems,frequently dominated by relative information about the environment,where main concerns such as occlusion,illumination,background,object deformation,and object class variations are commonplace.In order to show the importance of top view surveillance,a collaborative robotics framework has been presented.It can assist in the detection and tracking of multiple objects in top view surveillance.The framework consists of a smart robotic camera embedded with the visual processing unit.The existing pre-trained deep learning models named SSD and YOLO has been adopted for object detection and localization.The detection models are further combined with different tracking algorithms,including GOTURN,MEDIANFLOW,TLD,KCF,MIL,and BOOSTING.These algorithms,along with detection models,help to track and predict the trajectories of detected objects.The pre-trained models are employed;therefore,the generalization performance is also investigated through testing the models on various sequences of top view data set.The detection models achieved maximum True Detection Rate 93%to 90%with a maximum 0.6%False Detection Rate.The tracking results of different algorithms are nearly identical,with tracking accuracy ranging from 90%to 94%.Furthermore,a discussion has been carried out on output results along with future guidelines.展开更多
There are two main trends in the development of unmanned aerial vehicle(UAV)technologies:miniaturization and intellectualization,in which realizing object tracking capabilities for a nano-scale UAV is one of the most ...There are two main trends in the development of unmanned aerial vehicle(UAV)technologies:miniaturization and intellectualization,in which realizing object tracking capabilities for a nano-scale UAV is one of the most challenging problems.In this paper,we present a visual object tracking and servoing control system utilizing a tailor-made 38 g nano-scale quadrotor.A lightweight visual module is integrated to enable object tracking capabilities,and a micro positioning deck is mounted to provide accurate pose estimation.In order to be robust against object appearance variations,a novel object tracking algorithm,denoted by RMCTer,is proposed,which integrates a powerful short-term tracking module and an efficient long-term processing module.In particular,the long-term processing module can provide additional object information and modify the short-term tracking model in a timely manner.Furthermore,a positionbased visual servoing control method is proposed for the quadrotor,where an adaptive tracking controller is designed by leveraging backstepping and adaptive techniques.Stable and accurate object tracking is achieved even under disturbances.Experimental results are presented to demonstrate the high accuracy and stability of the whole tracking system.展开更多
Inspired by human behaviors, a robot object tracking model is proposed on the basis of visual attention mechanism, which is fit for the theory of topological perception. The model integrates the image-driven, bottom-u...Inspired by human behaviors, a robot object tracking model is proposed on the basis of visual attention mechanism, which is fit for the theory of topological perception. The model integrates the image-driven, bottom-up attention and the object-driven, top-down attention, whereas the previous attention model has mostly focused on either the bottom-up or top-down attention. By the bottom-up component, the whole scene is segmented into the ground region and the salient regions. Guided by top-down strategy which is achieved by a topological graph, the object regions are separated from the salient regions. The salient regions except the object regions are the barrier regions. In order to estimate the model, a mobile robot platform is developed, on which some experiments are implemented. The experimental results indicate that processing an image with a resolution of 752 × 480 pixels takes less than 200 ms and the object regions are unabridged. The analysis obtained by comparing the proposed model with the existing model demonstrates that the proposed model has some advantages in robot object tracking in terms of speed and efficiency.展开更多
基金funded by the Yangtze River Delta Science and Technology Innovation Community Joint Research Project(2023CSJGG1600)the Natural Science Foundation of Anhui Province(2208085MF173)Wuhu“ChiZhu Light”Major Science and Technology Project(2023ZD01,2023ZD03).
文摘As the number and complexity of sensors in autonomous vehicles continue to rise,multimodal fusionbased object detection algorithms are increasingly being used to detect 3D environmental information,significantly advancing the development of perception technology in autonomous driving.To further promote the development of fusion algorithms and improve detection performance,this paper discusses the advantages and recent advancements of multimodal fusion-based object detection algorithms.Starting fromsingle-modal sensor detection,the paper provides a detailed overview of typical sensors used in autonomous driving and introduces object detection methods based on images and point clouds.For image-based detection methods,they are categorized into monocular detection and binocular detection based on different input types.For point cloud-based detection methods,they are classified into projection-based,voxel-based,point cluster-based,pillar-based,and graph structure-based approaches based on the technical pathways for processing point cloud features.Additionally,multimodal fusion algorithms are divided into Camera-LiDAR fusion,Camera-Radar fusion,Camera-LiDAR-Radar fusion,and other sensor fusion methods based on the types of sensors involved.Furthermore,the paper identifies five key future research directions in this field,aiming to provide insights for researchers engaged in multimodal fusion-based object detection algorithms and to encourage broader attention to the research and application of multimodal fusion-based object detection.
基金supported in part by the"Pioneer"and"Leading Goose"R&D Program of Zhejiang under Grant No. 2024C01169the National Natural Science Foundation of China under Grant Nos. 62441238 and U2441240。
文摘Visual object tracking(VOT),aiming to track a target object in a continuous video,is a fundamental and critical task in computer vision.However,the reliance on third-party resources(e.g.,dataset)for training poses concealed threats to the security of VOT models.In this paper,we reveal that VOT models are vulnerable to a poison-only and targeted backdoor attack,where the adversary can achieve arbitrary tracking predictions by manipulating only part of the training data.Specifically,we first define and formulate three different variants of the targeted attacks:size-manipulation,trajectory-manipulation,and hybrid attacks.To implement these,we introduce Random Video Poisoning(RVP),a novel poison-only strategy that exploits temporal correlations within video data by poisoning entire video sequences.Extensive experiments demonstrate that RVP effectively injects controllable backdoors,enabling precise manipulation of tracking behavior upon trigger activation,while maintaining high performance on benign data,thus ensuring stealth.Our findings not only expose significant vulnerabilities but also highlight that the underlying principles could be adapted for beneficial uses,such as dataset watermarking for copyright protection.
基金funded by theNational Science and TechnologyCouncil(NSTC),Taiwan,under grant numbers NSTC 113-2634-F-A49-007 and NSTC 112-2634-F-A49-007.
文摘To improve small object detection and trajectory estimation from an aerial moving perspective,we propose the Aerial View Attention-PRB(AVA-PRB)model.AVA-PRB integrates two attention mechanisms—Coordinate Attention(CA)and the Convolutional Block Attention Module(CBAM)—to enhance detection accuracy.Additionally,Shape-IoU is employed as the loss function to refine localization precision.Our model further incorporates an adaptive feature fusion mechanism,which optimizes multi-scale object representation,ensuring robust tracking in complex aerial environments.We evaluate the performance of AVA-PRB on two benchmark datasets:Aerial Person Detection and VisDrone2019-Det.The model achieves 60.9%mAP@0.5 on the Aerial Person Detection dataset,and 51.2%mAP@0.5 on VisDrone2019-Det,demonstrating its effectiveness in aerial object detection.Beyond detection,we propose a novel trajectory estimation method that improves movement path prediction under aerial motion.Experimental results indicate that our approach reduces path deviation by up to 64%,effectively mitigating errors caused by rapid camera movements and background variations.By optimizing feature extraction and enhancing spatialtemporal coherence,our method significantly improves object tracking under aerial moving perspectives.This research addresses the limitations of fixed-camera tracking,enhancing flexibility and accuracy in aerial tracking applications.The proposed approach has broad potential for real-world applications,including surveillance,traffic monitoring,and environmental observation.
基金supported by the STI 2030-Major Projects(2022ZD0211900,2022ZD0211902)STI 2030-Major Projects(2021ZD0204500,2021ZD0204503)+1 种基金National Natural Science Foundation of China(32171461)National Key Research and Development Program of China(2023YFC3208303)。
文摘Advancements in animal behavior quantification methods have driven the development of computational ethology,enabling fully automated behavior analysis.Existing multianimal pose estimation workflows rely on tracking-bydetection frameworks for either bottom-up or top-down approaches,requiring retraining to accommodate diverse animal appearances.This study introduces InteBOMB,an integrated workflow that enhances top-down approaches by incorporating generic object tracking,eliminating the need for prior knowledge of target animals while maintaining broad generalizability.InteBOMB includes two key strategies for tracking and segmentation in laboratory environments and two techniques for pose estimation in natural settings.The“background enhancement”strategy optimizesforeground-backgroundcontrastiveloss,generating more discriminative correlation maps.The“online proofreading”strategy stores human-in-the-loop long-term memory and dynamic short-term memory,enabling adaptive updates to object visual features.The“automated labeling suggestion”technique reuses the visual features saved during tracking to identify representative frames for training set labeling.Additionally,the“joint behavior analysis”technique integrates these features with multimodal data,expanding the latent space for behavior classification and clustering.To evaluate the framework,six datasets of mice and six datasets of nonhuman primates were compiled,covering laboratory and natural scenes.Benchmarking results demonstrated a24%improvement in zero-shot generic tracking and a 21%enhancement in joint latent space performance across datasets,highlighting the effectiveness of this approach in robust,generalizable behavior analysis.
基金supported by the National Natural Science Foundation of China under Grant 62177029the Postgraduate Research&Practice Innovation Program of Jiangsu Province(KYCX21_0740),China.
文摘Visual object tracking plays a crucial role in computer vision.In recent years,researchers have proposed various methods to achieve high-performance object tracking.Among these,methods based on Transformers have become a research hotspot due to their ability to globally model and contextualize information.However,current Transformer-based object tracking methods still face challenges such as low tracking accuracy and the presence of redundant feature information.In this paper,we introduce self-calibration multi-head self-attention Transformer(SMSTracker)as a solution to these challenges.It employs a hybrid tensor decomposition self-organizing multihead self-attention transformermechanism,which not only compresses and accelerates Transformer operations but also significantly reduces redundant data,thereby enhancing the accuracy and efficiency of tracking.Additionally,we introduce a self-calibration attention fusion block to resolve common issues of attention ambiguities and inconsistencies found in traditional trackingmethods,ensuring the stability and reliability of tracking performance across various scenarios.By integrating a hybrid tensor decomposition approach with a self-organizingmulti-head self-attentive transformer mechanism,SMSTracker enhances the efficiency and accuracy of the tracking process.Experimental results show that SMSTracker achieves competitive performance in visual object tracking,promising more robust and efficient tracking systems,demonstrating its potential to providemore robust and efficient tracking solutions in real-world applications.
基金supported in part by National Natural Science Foundation of China(No.62176041)in part by Excellent Science and Technique Talent Foundation of Dalian(No.2022RY21).
文摘Significant advancements have beenwitnessed in visual tracking applications leveragingViT in recent years,mainly due to the formidablemodeling capabilities of Vision Transformer(ViT).However,the strong performance of such trackers heavily relies on ViT models pretrained for long periods,limitingmore flexible model designs for tracking tasks.To address this issue,we propose an efficient unsupervised ViT pretraining method for the tracking task based on masked autoencoders,called TrackMAE.During pretraining,we employ two shared-parameter ViTs,serving as the appearance encoder and motion encoder,respectively.The appearance encoder encodes randomly masked image data,while the motion encoder encodes randomly masked pairs of video frames.Subsequently,an appearance decoder and a motion decoder separately reconstruct the original image data and video frame data at the pixel level.In this way,ViT learns to understand both the appearance of images and the motion between video frames simultaneously.Experimental results demonstrate that ViT-Base and ViT-Large models,pretrained with TrackMAE and combined with a simple tracking head,achieve state-of-the-art(SOTA)performance without additional design.Moreover,compared to the currently popular MAE pretraining methods,TrackMAE consumes only 1/5 of the training time,which will facilitate the customization of diverse models for tracking.For instance,we additionally customize a lightweight ViT-XS,which achieves SOTA efficient tracking performance.
基金the National Natural Science Foundation of China (No.61627810)the National Science and Technology Major Program of China (No.2018YFB1305003)the National Defense Science and Technology Outstanding Youth Science Foundation (No.2017-JCJQ-ZQ-031)。
文摘In order to solve the problem of small object size and low detection accuracy under the unmanned aerial vehicle(UAV)platform,the object detection algorithm based on deep aggregation network and high-resolution fusion module is studied.Furthermore,a joint network of object detection and feature extraction is studied to construct a real-time multi-object tracking algorithm.For the problem of object association failure caused by UAV movement,image registration is applied to multi-object tracking and a camera motion discrimination model is proposed to improve the speed of the multi-object tracking algorithm.The simulation results show that the algorithm proposed in this study can improve the accuracy of multi-object tracking under the UAV platform,and effectively solve the problem of association failure caused by UAV movement.
基金supported by the New Cornerstone Science Foundation through the XPLORER PRIZE,China(No.XPLORER-2024-1036)the independent research project of the National Key Laboratory of Strength and Structural Integrity,China(No.BYST-QZSYS-24-072-5)。
文摘Crack length measurement algorithms based on computer vision have shown promising engineering application prospects in the field of aircraft fatigue crack monitoring.However,due to the complexity of the monitoring environment,the subtle visual features of small fatigue cracks,and the impact of structural elastic deformation,directly applying object segmentation algorithms often results in significant measurement errors.Therefore,this paper proposes a high-precision crack length measurement method based on Bidirectional Target Tracking Model(Bi2TM),which integrates crack tip localization,interference identification,and length compensation.First,a general object segmentation model is used to perform rough crack segmentation.Then,the Bi2TM network,combined with the visual features of the structure in different stress states,is employed to track the bidirectional position of the crack tip in the“open”and“closed”states.This ultimately enables interference identification within the rough segmented crack region,achieving highprecision length measurement.In a high-interference environment of aircraft fatigue testing,the proposed method is used to measure 1000 crack images ranging from 1 mm to 11 mm.For more than 90%of the samples,the measurement error is less than 5 pixels,demonstrating significant advantages over the existing methods.
基金supported by the National Natural Science Foundation of China(GrantNo.62302086)the Natural Science Foundation of Liaoning Province(Grant No.2023-MSBA-070)the Fundamental Research Funds for the Central Universities(Grant No.N2317005).
文摘Multi-modal 3D object detection has achieved remarkable progress,but it is often limited in practical industrial production because of its high cost and low efficiency.The multi-view camera-based method provides a feasible solution due to its low cost.However,camera data lacks geometric depth,and only using camera data to obtain high accuracy is challenging.This paper proposes a multi-modal Bird-Eye-View(BEV)distillation framework(MMDistill)to make a trade-off between them.MMDistill is a carefully crafted two-stage distillation framework based on teacher and student models for learning cross-modal knowledge and generating multi-modal features.It can improve the performance of unimodal detectors without introducing additional costs during inference.Specifically,our method can effectively solve the cross-gap caused by the heterogeneity between data.Furthermore,we further propose a Light Detection and Ranging(LiDAR)-guided geometric compensation module,which can assist the student model in obtaining effective geometric features and reduce the gap between different modalities.Our proposed method generally requires fewer computational resources and faster inference speed than traditional multi-modal models.This advancement enables multi-modal technology to be applied more widely in practical scenarios.Through experiments,we validate the effectiveness and superiority of MMDistill on the nuScenes dataset,achieving an improvement of 4.1%mean Average Precision(mAP)and 4.6%NuScenes Detection Score(NDS)over the baseline detector.In addition,we also present detailed ablation studies to validate our method.
基金the National Natural Science Foundation of China (No.61673269)。
文摘Visual object tracking is an important issue that has received long-term attention in computer vision.The ability to effectively handle occlusion,especially severe occlusion,is an important aspect of evaluating the performance of object tracking algorithms in long-term tracking,and is of great significance to improving the robustness of object tracking algorithms.However,most object tracking algorithms lack a processing mechanism specifically for occlusion.In the case of occlusion,due to the lack of target information,it is necessary to predict the target position based on the motion trajectory.Kalman filtering and particle filtering can effectively predict the target motion state based on the historical motion information.A single object tracking method,called probabilistic discriminative model prediction(PrDiMP),is based on the spatial attention mechanism in complex scenes and occlusions.In order to improve the performance of PrDiMP,Kalman filtering,particle filtering and linear filtering are introduced.First,for the occlusion situation,Kalman filtering and particle filtering are respectively introduced to predict the object position,thereby replacing the detection result of the original tracking algorithm and stopping recursion of target model.Second,for detection-jump problem of similar objects in complex scenes,a linear filtering window is added.The evaluation results on the three datasets,including GOT-10k,UAV123 and LaSOT,and the visualization results on several videos,show that our algorithms have improved tracking performance under occlusion and the detection-jump is effectively suppressed.
文摘Visual object-tracking is a fundamental task applied in many applications of computer vision. Particle filter is one of the techniques which has been widely used in object tracking. Due to the virtue of extendability and flexibility on both linear and non-linear environments, various particle filter-based trackers have been proposed in the literature. However, the conventional approach cannot handle very large videos efficiently in the current data intensive information age. In this work, a parallelized particle filter is provided in a distributed framework provided by the Hadoop/Map-Reduce infrastructure to tackle object-tracking tasks. The experiments indicate that the proposed algorithm has a better convergence and accuracy as compared to the traditional particle filter. The computational power and the scalability of the proposed particle filter in single object tracking have been enhanced as well.
文摘Droplet microfluidics,which encapsulates individual cells within separate microreactors,has become an essential tool for single-cell phenotypic and genotypic analysis.However,the efficiency of single-cell encapsulation is limited by the Poisson distribution governing the encapsulation process,resulting in most droplets being either empty or containing multiple cells.Traditional single-cell sorting methods typically rely on fluorescence labeling for identification,but this approach not only increases experimental costs and complexity but can also impact cell viability.Additionally,current label-free sorting methods still encounter difficulties in accurately detecting multicellular droplets and small cellular aggregates.To address these challenges,this paper proposes an intelligent sorting system that combines YOLOv8 object detection and BoTSORT tracking algorithms.This system enables real-time analysis of droplet images,facilitating precise identification,counting,and automated sorting of target droplets.To validate the system’s performance,polystyrene microspheres were used to simulate real cells in sorting tests.The results demonstrated that,under label-free conditions,the system significantly outperformed traditional fluorescence labeling methods in both classification accuracy and sorting efficiency.This system provides an effective,label-free solution for cell sorting,with potential applications in precision medicine,single-cell sequencing,and drug screening.
文摘An improved estimation of motion vectors of feature points is proposed for tracking moving objects of dynamic image sequence. Feature points are firstly extracted by the improved minimum intensity change （MIC） algorithm. The matching points of these feature points are then determined by adaptive rood pattern searching. Based on the random sample consensus （RANSAC） method, the background motion is finally compensated by the parameters of an affine transform of the background motion. With reasonable morphological filtering, the moving objects are completely extracted from the background, and then tracked accurately. Experimental results show that the improved method is successful on the motion background compensation and offers great promise in tracking moving objects of the dynamic image sequence.
基金supported by the National Natural Science Foundation of China(Grant No.51009040)Heilongjiang Postdoctoral Fund(Grant No.LBH-Z11205)+1 种基金the National High Technology Research and Development Program of China(863 Program,Grant No.2011AA09A106)the China Postdoctoral Science Foundation(Grant No.2012M510928)
文摘This paper describes a new framework for object detection and tracking of AUV including underwater acoustic data interpolation, underwater acoustic images segmentation and underwater objects tracking. This framework is applied to the design of vision-based method for AUV based on the forward looking sonar sensor. First, the real-time data flow （underwater acoustic images） is pre-processed to form the whole underwater acoustic image, and the relevant position information of objects is extracted and determined. An improved method of double threshold segmentation is proposed to resolve the problem that the threshold cannot be adjusted adaptively in the traditional method. Second, a representation of region information is created in light of the Gaussian particle filter. The weighted integration strategy combining the area and invariant moment is proposed to perfect the weight of particles and to enhance the tracking robustness. Results obtained on the real acoustic vision platform of AUV during sea trials are displayed and discussed. They show that the proposed method can detect and track the moving objects underwater online, and it is effective and robust.
基金supported by the National Natural Science Foundation of China(No.61671002)
文摘This paper addresses the problem of real-time object tracking for unmanned aerial vehicles. We consider the task of object tracking as a classification problem. Training a good classifier always needs a huge number of samples, which is always time-consuming and not suitable for realtime applications. In this paper, we transform the large-scale least-squares problem in the spatial domain to a series of small-scale least-squares problems with constraints in the Fourier domain using the correlation filter technique. Then, this problem is efficiently solved by two stages. In the first stage, a fast method based on recursive least squares is used to solve the correlation filter problem without constraints in the Fourier domain. In the second stage, a weight matrix is constructed to prune the solution attained in the first stage to approach the constraints in the spatial domain. Then, the pruned classifier is used for tracking. To evaluate proposed tracker’s performance, comprehensive experiments are conducted on challenging aerial sequences in the UAV123 dataset. Experimental results demonstrate that proposed approach achieves a state-ofthe-art tracking performance in aerial sequences and operates at a mean speed of beyond 40 frames/s. For further analysis of proposed tracker’s robustness, extensive experiments are also performed on recent benchmarks OTB50, OTB100, and VOT2016.
文摘A method for moving object recognition and tracking in the intelligent traffic monitoring system is presented. For the shortcomings and deficiencies of the frame-subtraction method, a redundant discrete wavelet transform （RDWT） based moving object recognition algorithm is put forward, which directly detects moving objects in the redundant discrete wavelet transform domain. An improved adaptive mean-shift algorithm is used to track the moving object in the follow up frames. Experimental results show that the algorithm can effectively extract the moving object, even though the object is similar to the background, and the results are better than the traditional frame-subtraction method. The object tracking is accurate without the impact of changes in the size of the object. Therefore the algorithm has a certain practical value and prospect.
文摘为了解决现有的目标跟踪算法在煤矿复杂环境下存在精度低和实时性差的问题,基于Tracking by Detection(TBD)范式,提出了YOLO-FasterNet+ByteTrack的煤矿人员跟踪算法。首先,构建FasterNet-Block特征提取模块改进YOLOv7的Backbone,提升目标检测阶段的实时性;然后,通过在Neck中引入CBAM注意力机制,提升模型在复杂场景下的特征感知能力;接着,在目标检测的解码阶段引入Soft-NMS,优化模型在人员交叠场景下的检测精度;最后,在目标跟踪阶段,针对人员重叠和遮挡导致的目标ID翻转问题,设计了一种融合GRU和卡尔曼滤波的多目标运动特征预测机制,有效提升了煤矿人员跟踪的准确性。实验结果表明:YOLOFasterNet在煤矿人员数据集上相对于YOLOv7的平均精度提高了3.6%,检测速度提升了8.2FPS;在自定义跟踪数据集GBMOT上,所提目标跟踪算法相对于ByteTrack,MOTA值提升了1.7%,IDSW减少了149次。
基金the Framework of International Cooperation Program managed by the National Research Foundation of Korea(2019K1A3A1A8011295711).
文摘Collaborative Robotics is one of the high-interest research topics in the area of academia and industry.It has been progressively utilized in numerous applications,particularly in intelligent surveillance systems.It allows the deployment of smart cameras or optical sensors with computer vision techniques,which may serve in several object detection and tracking tasks.These tasks have been considered challenging and high-level perceptual problems,frequently dominated by relative information about the environment,where main concerns such as occlusion,illumination,background,object deformation,and object class variations are commonplace.In order to show the importance of top view surveillance,a collaborative robotics framework has been presented.It can assist in the detection and tracking of multiple objects in top view surveillance.The framework consists of a smart robotic camera embedded with the visual processing unit.The existing pre-trained deep learning models named SSD and YOLO has been adopted for object detection and localization.The detection models are further combined with different tracking algorithms,including GOTURN,MEDIANFLOW,TLD,KCF,MIL,and BOOSTING.These algorithms,along with detection models,help to track and predict the trajectories of detected objects.The pre-trained models are employed;therefore,the generalization performance is also investigated through testing the models on various sequences of top view data set.The detection models achieved maximum True Detection Rate 93%to 90%with a maximum 0.6%False Detection Rate.The tracking results of different algorithms are nearly identical,with tracking accuracy ranging from 90%to 94%.Furthermore,a discussion has been carried out on output results along with future guidelines.
基金supported in part by the Institute for Guo Qiang of Tsinghua University(2019GQG1023)in part by Graduate Education and Teaching Reform Project of Tsinghua University(202007J007)+1 种基金in part by National Natural Science Foundation of China(U19B2029,62073028,61803222)in part by the Independent Research Program of Tsinghua University(2018Z05JDX002)。
文摘There are two main trends in the development of unmanned aerial vehicle(UAV)technologies:miniaturization and intellectualization,in which realizing object tracking capabilities for a nano-scale UAV is one of the most challenging problems.In this paper,we present a visual object tracking and servoing control system utilizing a tailor-made 38 g nano-scale quadrotor.A lightweight visual module is integrated to enable object tracking capabilities,and a micro positioning deck is mounted to provide accurate pose estimation.In order to be robust against object appearance variations,a novel object tracking algorithm,denoted by RMCTer,is proposed,which integrates a powerful short-term tracking module and an efficient long-term processing module.In particular,the long-term processing module can provide additional object information and modify the short-term tracking model in a timely manner.Furthermore,a positionbased visual servoing control method is proposed for the quadrotor,where an adaptive tracking controller is designed by leveraging backstepping and adaptive techniques.Stable and accurate object tracking is achieved even under disturbances.Experimental results are presented to demonstrate the high accuracy and stability of the whole tracking system.
基金supported by National Basic Research Program of China (973 Program) (No. 2006CB300407)National Natural Science Foundation of China (No. 50775017)
文摘Inspired by human behaviors, a robot object tracking model is proposed on the basis of visual attention mechanism, which is fit for the theory of topological perception. The model integrates the image-driven, bottom-up attention and the object-driven, top-down attention, whereas the previous attention model has mostly focused on either the bottom-up or top-down attention. By the bottom-up component, the whole scene is segmented into the ground region and the salient regions. Guided by top-down strategy which is achieved by a topological graph, the object regions are separated from the salient regions. The salient regions except the object regions are the barrier regions. In order to estimate the model, a mobile robot platform is developed, on which some experiments are implemented. The experimental results indicate that processing an image with a resolution of 752 × 480 pixels takes less than 200 ms and the object regions are unabridged. The analysis obtained by comparing the proposed model with the existing model demonstrates that the proposed model has some advantages in robot object tracking in terms of speed and efficiency.
