Object Tracking Research Articles

A Quantum Evolutionary Learning Tracker for Video

Video object tracking has been a popular area in the field of computer vision. As video data evolves, more special perspectives and challenging video data are constantly kept up to date. This poses challenges for object tracking tasks and places higher demands on the generalization capabilities of the models. In this paper, we propose a novel quantum evolutionary learning tracker for video. The model combines quantum evolution with deep networks for tracking video objects. The model uses a quantum evolutionary learning tracker to generate a reliable population of candidate regions and a deep network for classification. In particular, the quantum evolutionary predictor predicts the object motion state through rotation operator and trajectory inference, and provides motion state information for the tracker. The predictor can incorporate object history contextual information and can provide stable candidate estimation populations for the model in case of failure of appearance features. Both quantum evolution and deep networks are combined to form an end-to-end online video object tracker. In addition, we propose a new video object tracking evaluation algorithm, Balanced Intersection over Union. The evaluation algorithm uses aspect ratios to balance the share of overlap and distance. Finally, we test the model on the OTB 2015 dataset for natural video and on the SV248A10-SOT dataset for satellite video. The performance of the proposed model is also analyzed and validated by comparing it with more than twenty classical tracker models. The experimental results show that our model has high generalization ability and robustness.

Real-Time Satellite Constellation Scheduling for Event-Triggered Cooperative Tracking of Space Objects

Real-Time Satellite Constellation Scheduling for Event-Triggered Cooperative Tracking of Space Objects

Image convolution techniques integrated with YOLOv3 algorithm in motion object data filtering and detection.

In order to address the challenges of identifying, detecting, and tracking moving objects in video surveillance, this paper emphasizes image-based dynamic entity detection. It delves into the complexities of numerous moving objects, dense targets, and intricate backgrounds. Leveraging the You Only Look Once (YOLOv3) algorithm framework, this paper proposes improvements in image segmentation and data filtering to address these challenges. These enhancements form a novel multi-object detection algorithm based on an improved YOLOv3 framework, specifically designed for video applications. Experimental validation demonstrates the feasibility of this algorithm, with success rates exceeding 60% for videos such as "jogging", "subway", "video 1", and "video 2". Notably, the detection success rates for "jogging" and "video 1" consistently surpass 80%, indicating outstanding detection performance. Although the accuracy slightly decreases for "Bolt" and "Walking2", success rates still hover around 70%. Comparative analysis with other algorithms reveals that this method's tracking accuracy surpasses that of particle filters, Discriminative Scale Space Tracker (DSST), and Scale Adaptive Multiple Features (SAMF) algorithms, with an accuracy of 0.822. This indicates superior overall performance in target tracking. Therefore, the improved YOLOv3-based multi-object detection and tracking algorithm demonstrates robust filtering and detection capabilities in noise-resistant experiments, making it highly suitable for various detection tasks in practical applications. It can address inherent limitations such as missed detections, false positives, and imprecise localization. These improvements significantly enhance the efficiency and accuracy of target detection, providing valuable insights for researchers in the field of object detection, tracking, and recognition in video surveillance.

228 Expert Hand Motion During Microvascular Anastomosis Simulation: Big Data Analysis Using Machine Learning Hand Detection

INTRODUCTION: Microanastomosis is a complex surgical technique that demands delicate and coordinated hand movement. Advances in machine learning technology have enabled new methods to study human motion. Convolutional neural networks are algorithms specialized in visual tasks that enable discrete object detection and tracking including the hands of a surgeon. METHODS: A machine learning hand detector was developed using a pre-trained convolutional neural network capable of tracking 21 hand landmarks without sensors attached. En-to-side microanastomosis was simulated using 2 mm polyvinyl-alcohol vessels and 10/0 micro-suture and recorded with a video camera capturing hand movement in the surgical field. Participants (n = 4) were expert cerebrovascular surgeons with significant operative experience in cerebral bypass surgery. Calculation of motion of 14 hand joint angles was performed using a vectorization method from landmark detections. Time series analysis was performed to calculate the absolute angular motion for each joint of the hand. RESULTS: Four minutes of simulation for each surgeon yielded a total of 784.294 angular detections (14 hand joints total). For the dominant hand, the mean (SD) absolute angular motion was expert 1: 178.2 rad (98.3); expert 2: 261.8 rad (164.3); expert 3: 449.9 rad (307); expert 4: 380.1 rad (198.1).For the non-dominant hand, the mean (SD) absolute angular motion was expert 1: 132.9 rad (64.6), expert 2: 149.2 rad (72.3); expert 3: 389.2 rad (243); expert 4: 162.6 rad (87.2). CONCLUSIONS: A machine learning hand detector system can calculate hand joint motion during microvascular anastomosis. The absolute angular motion is a quantitative performance metric providing information about hand stability during microanastomosis simulation. Such studies using powerful artificial intelligence detection algorithms may have educational potential in the assessment of complex and delicate surgical skills and supervision of surgical training.

Accident Avoidance using Railway Track Detection System using IoT

An IoT-based railway track crack and object discovery framework is an innovative arrangement for guaranteeing the security of railroad frameworks. The framework employs a combination of sensors and cameras that are introduced along the railroad track to screen it for any signs of harm or hindrances. The sensors can identify vibrations and changes within the track, which can show the nearness of a break or other sort of harm. The cameras can capture pictures and video of the track, such as fallen branches, flotsam, and jetsam, or indeed individuals or creatures. Once the framework recognizes an issue, it can send an alarm to the railroad administrators or upkeep workforce, permitting them to rapidly address the issue sometime soon if it causes any mischances or delays. An IOT-based railroad track break and protest location framework can incredibly move forward the security and proficiency of railroad operations. Key Words: Embedded system, Internet of things, ultrasonic sensor, IR detector, Arduino Uno, GPS

Chicken tracking for location mapping of lameness chickens using YOLOv8 and deep learning-based tracking algorithm

The chicken farming industry is one of the biggest food industries that supports the achievement of food security internationally. Farmers need an independent tool that can monitor the welfare conditions of chickens in cages. Using their tools, farmers can ideally detect the condition of chickens. Lameness chickens, can be known for activity and dredging of their location in the cage. Occlusion, and background in the cage are interesting challenges. By observing behavior, image handling practices can be used to identify tainted chicks and provide an early warning of sickness in chickens. In this study, you only look once, version 8 (YOLOv8) which is a convolutional neural network (CNN) network model was chosen to perform the detection, tracking, and mapping of chicken locations. YOLOv8 was combined with various algorithm optimizers to improve training performance, such as root mean square (RMS) Prop, stochastic gradient descent (SGD), ADAM, and ADAMW. Multi-object tracking algorithms such as BOT-sort and ByteTrack are also used to improve tracking performance. Based on the results, YOLOv8 with combinations of optimizer algorithms ADAMW has the best mAP, support, precision and F1-score values compared to the others, with 0.936, 0.993, 0.990, 0.991. Meanwhile, for multi object tracking, ByteTrack is faster in inference time(s) values compared to the others, with 0.2.

Anomaly Detection in Crowd Surveillance using Edge Computing: A Comprehensive Survey

Abstract: As urban environments continue to witness the proliferation of surveillance systems, ensuring public safety and security has become an increasingly challenging endeavor. Anomaly detection in crowded areas has emerged as a crucial task in these settings, aimed at identifying suspicious activities or potential threats. To meet the demands of real-time monitoring and analysis, edge computing has gained prominence as a critical technology. This survey paper provides a comprehensive overview of the state-of-the-art in anomaly detection, its use in crowd surviellance and the role played by edge computing in anomaly detection .The paper reviews an extensive body of literature encompassing various techniques and methodologies employed for anomaly detection in crowded scenes. It explores the evolution of traditional video-based approaches, such as motion analysis and object tracking, and the recent advancements leveraging deep learning, including various models in machine learning. These techniques are examined for their applicability in crowd surveillance scenarios.Various commonly used datasets to measure the quality of anomaly detection are also explored, along with their attributes and descriptions.The survey analyzes how edge computing solutions, such as edge AI accelerators and edge devices, enable faster and more context-aware processing of video data, while also addressing issues related to bandwidth constraints, privacy concerns, and scalability; paving way for a further research in harnessing the power of edge computing in crowd surveillance anomaly detection.

A Study of YOLO-Based Object Detection for Visually Impaired Individuals

Abstract: An essential area of investigation involves recognizing and tracking objects, given the frequent changes in object motion, scene dimensions, occlusions, appearance, ego-motion, and lighting variations. Effective object tracking heavily relies on feature selection due to its critical role. This process is integral to numerous real-time applications like vehicle detection and video surveillance. To address detection challenges, tracking methods need to be adept at handling object movement and appearance changes. Among these methods, tracking algorithms play a pivotal role in smoothing video sequences and receive significant attention. However, only a few approaches leverage previously gathered data on object attributes such as shape and color. This study delves into a tracking algorithm that encompasses all these object properties. Its aim is to investigate and assess past methodologies for tracking and detecting moving objects across various stages of video sequences. Additionally, it aims to identify existing challenges and propose innovative strategies to enhance object tracking throughout video frames

Intelligent Traffic System: Yolov8 and Deepsort in Car Detection, Tracking and Counting

Abstract: As the number of vehicles on the road increases, it becomes essential to quickly and accurately identify them to better detect traffic congestion and improve traffic management. DeepSORT (simple online and real-time tracking with a deep association metric) multitarget tracking algorithm in vehicle tracking. Due to the strong dependence of the DeepSORT algorithm on target detection, a YOLOv8 vehicle detection algorithm was proposed based on YOLOv7, which provides accurate and fast vehicle detection data to the DeepSORT algorithm. DeepSORT introduces deep learning into the SORT algorithm by adding an appearance descriptor to reduce identity switches, making tracking more efficient. DeepSORT uses a better association metric that combines both motion and appearance descriptors. DeepSORT can be defined as the tracking algorithm that tracks objects not only based on the velocity and motion of the object but also the appearance of the object. This vehicle detection also uses the DeepSORT algorithm to help count the number of vehicles that pass in the video effectively. YOLOv8 and the DeepSORT algorithm collaborate to identify and follow vehicles creating a model, for vehicle recognition that showcases their ability to track cars efficiently. From this paper, the model Yolov8 has achieved state-of-the-art results

Development of Augmented Reality Application as An Educational Media for Visitors to Museum Pusaka Keraton Kasepuhan Cirebon Using Object Tracking Method and Fast Corner Detection Algorithm Based on Android

Augmented Reality is a technology that combines two-dimensional or three-dimensional virtual objects and projects these virtual objects in real time. One implementation of Augmented Reality in the tourism sector is to educate museum visitors. Museum Pusaka Keraton Kasepuhan Cirebon does not yet have a touch of technology to attract visitors, and the public paradigm is that visiting the museum only sees heirloom objects, nothing interesting or unique. The aim of the research carried out by the author is to apply Augmented Reality with the Object Tracking method and the FAST Corner Detection algorithm to educate museum visitors, so that it can attract visitors' attention. By utilizing these methods and algorithms, it can be easier for visitors to explore heirloom objects to obtain the desired information. So the results obtained from the research conducted by the author are that the response time for objects appearing using the Tracking Object method and the FAST Corner Detection algorithm in environments that use glass is an average of 1.52 seconds to 2.40 seconds and that does not use glass, namely 2.84 seconds to 4.71 seconds with a level of confidence at the 95% level.

Multi-camera multi-object tracking on the move via single-stage global association approach

The development of autonomous vehicles generates a tremendous demand for a low-cost solution based on a complete set of camera sensors to perceive the environment around the car. Towards this solution, it is essential for object detection and tracking to address new challenges in multi-camera settings. To address these challenges, this work introduces novel Single-Stage Global Association Tracking approaches to associate one or more detections from multi-cameras with tracked objects. These approaches aim to solve fragment-tracking issues caused by inconsistent 3D object detection. Moreover, our models also improve the detection accuracy of the standard vision-based 3D object detectors in the nuScenes detection challenge. The extensive experimental results on the nuScenes dataset demonstrate the benefits of the proposed method, which outperforms prior vision-based tracking methods in multi-camera settings.

Dynamic region-aware transformer backbone network for visual tracking

In visual tracking, the Transformer architecture is widely used because it can capture the global dependencies of sequence data without inductive bias. However, the attention mechanism of Transformer will bring ultra-high computational complexity and space occupancy, so that the tracking task cannot meet the real-time requirements. In this paper, we explore a sparsity region-aware attention mechanism. The sparse attention mechanism retains the regions with semantic relevance, and performs fine-grained attention calculation in this region. In the region-aware attention mechanism, a DropKey technique is introduced to reduce model over-fitting and improve the generalization ability of the model. Using region-aware attention as the basic building block, we design a dynamic region-aware Transformer backbone for visual tracking. This backbone network can effectively reduce the computational complexity while exploring global context dependencies. Based on the region-aware Transformer backbone network, this paper proposes a dynamic region-aware Transformer backbone visual tracking algorithm, which uses an optimization based model predictor to fully fuse object appearance and background information, so as to achieve more robust object tracking. The proposed tracker is trained in an end-to-end manner and experimentally evaluated on eight tracking benchmarks. Experimental results show that the algorithm has good tracking performance, especially in the application of unmanned aerial vehicle (UAV) tracking, our proposed tracker achieves an area under curve (AUC) score of 66.5% on the UAV123 dataset. Code is available at https://github.com/YSGFF/RTDiMP.

NeXtFusion: Attention-Based Camera-Radar Fusion Network for Improved Three-Dimensional Object Detection and Tracking

Accurate perception is crucial for autonomous vehicles (AVs) to navigate safely, especially in adverse weather and lighting conditions where single-sensor networks (e.g., cameras or radar) struggle with reduced maneuverability and unrecognizable targets. Deep Camera-Radar fusion neural networks offer a promising solution for reliable AV perception under any weather and lighting conditions. Cameras provide rich semantic information, while radars act like an X-ray vision, piercing through fog and darkness. This work proposes a novel, efficient Camera-Radar fusion network called NeXtFusion for robust AV perception with an improvement in object detection accuracy and tracking. Our proposed approach of utilizing an attention module enhances crucial feature representation for object detection while minimizing information loss from multi-modal data. Extensive experiments on the challenging nuScenes dataset demonstrate NeXtFusion’s superior performance in detecting small and distant objects compared to other methods. Notably, NeXtFusion achieves the highest mAP score (0.473) on the nuScenes validation set, outperforming competitors like OFT (35.1% improvement) and MonoDIS (9.5% improvement). Additionally, NeXtFusion demonstrates strong performance in other metrics like mATE (0.449) and mAOE (0.534), highlighting its overall effectiveness in 3D object detection. Furthermore, visualizations of nuScenes data processed by NeXtFusion further demonstrate its capability to handle diverse real-world scenarios. These results suggest that NeXtFusion is a promising deep fusion network for improving AV perception and safety for autonomous driving.

Multimodal Features Alignment for Vision–Language Object Tracking

Vision–language tracking presents a crucial challenge in multimodal object tracking. Integrating language features and visual features can enhance target localization and improve the stability and accuracy of the tracking process. However, most existing fusion models in vision–language trackers simply concatenate visual and linguistic features without considering their semantic relationships. Such methods fail to distinguish the target’s appearance features from the background, particularly when the target changes dramatically. To address these limitations, we introduce an innovative technique known as multimodal features alignment (MFA) for vision–language tracking. In contrast to basic concatenation methods, our approach employs a factorized bilinear pooling method that conducts squeezing and expanding operations to create a unified feature representation from visual and linguistic features. Moreover, we integrate the co-attention mechanism twice to derive varied weights for the search region, ensuring that higher weights are placed on the aligned visual and linguistic features. Subsequently, the fused feature map with diverse distributed weights serves as the search region during the tracking phase, facilitating anchor-free grounding to predict the target’s location. Extensive experiments are conducted on multiple public datasets, and our proposed tracker obtains a success score of 0.654/0.553/0.447 and a precision score of 0.872/0.556/0.513 on OTB-LANG/LaSOT/TNL2K. These results are satisfying compared with those of recent state-of-the-art vision–language trackers.

In defense and revival of Bayesian filtering for thermal infrared object tracking

Deep learning-based methods monopolize the latest research in the field of thermal infrared (TIR) object tracking. However, relying solely on deep learning models to obtain better tracking results requires carefully selecting feature information that is beneficial to representing the target object and designing a reasonable template update strategy, which undoubtedly increases the difficulty of model design. Thus, recent TIR tracking methods face many challenges in complex scenarios. This paper introduces a novel Deep Bayesian Filtering (DBF) method to enhance TIR tracking in these challenging situations. DBF is distinctive in its dual-model structure: the system and observation models. The system model leverages motion data to estimate the potential positions of the target object based on two-dimensional Brownian motion, thus generating a prior probability. Following this, the observation model comes into play upon capturing the TIR image. It serves as a classifier and employs infrared information to ascertain the likelihood of these estimated positions, creating a likelihood probability. According to the guidance of the two models, the position of the target object can be determined, and the template can be dynamically updated. Experimental analysis across several benchmark datasets reveals that DBF achieves competitive performance, surpassing most existing TIR tracking methods in complex scenarios.

Sensor and sensor fusion technology in autonomous vehicles

The perception and navigation of autonomous vehicles heavily rely on the utilization of sensor technology and the integration of sensor fusion techniques, which play an essential role in ensuring a secure and proficient understanding of the vehicle's environment.This paper highlights the significance of sensors in autonomous vehicles and how sensor fusion techniques enhance their capabilities. Firstly, the paper introduces the different types of sensors commonly used in autonomous vehicles and explains their principles of operation, strengths, and limitations in capturing essential information about the vehicles environment. Next, the paper discusses various sensor fusion algorithms, such as Kalman filters and particle filters. Furthermore, the paper explores the challenges associated with sensor fusion and addresses the issue of handling sensor failures or uncertainties. The benefits of sensor fusion technology in autonomous vehicles are also presented. These include improved perception of the environment, enhanced object recognition and tracking, better trajectory planning, and enhanced safety through redundancy and fault tolerance. Lastly, the paper discusses the advancements and highlights the integration of artificial intelligence and machine learning techniques to optimize sensor fusion algorithms and improve the overall autonomy of the vehicle. Following thorough analysis, the deduction can be made that sensor and sensor fusion technology assume a critical function in facilitating efficient and secure autonomous vehicle navigation within intricate surroundings.

Observer-based finite-time control for trajectory tracking of wheeled mobile robots with kinematic disturbances

Wheeled Mobile Robots (WMRs) are systems with applications in diverse fields such as transportation, civilian services, military use, and space exploration. Then, their use will continue increasing, making WMRs an essential research topic that deserves further study. To this end, this work presents a novel observer-based finite-time controller for trajectory tracking control of WMRs disturbed by kinematic disturbances. In the proposed approach, the kinematic model of the WMR is transformed into a set of two decoupled second-order systems. Then, the proposed controller is divided into two parts. The first one employs an observer to estimate the effect of the kinematic disturbances. The second part consists of a finite-time controller designed to achieve finite-time convergence of the tracking error. A detailed synthesis procedure theoretically demonstrates the feasibility of the proposed controller. Subsequently, the proposed scheme is compared against finite-time, feedback, and H∞ controllers. Exhaustive numerical simulations show that the proposed new control methodology achieves the trajectory tracking objective despite kinematic disturbances and outperforms the other control procedures. Finally, some comments and numerical results are given to clarify how the proposed control methodology can be used to design new controllers for trajectory tracking in WMRs and demonstrate that the new proposal remains to have a good performance when the system’s coordinates are corrupted by measurement noise.

Object Detection and Tracking with YOLO and the Sliding Innovation Filter.

Object detection and tracking are pivotal tasks in machine learning, particularly within the domain of computer vision technologies. Despite significant advancements in object detection frameworks, challenges persist in real-world tracking scenarios, including object interactions, occlusions, and background interference. Many algorithms have been proposed to carry out such tasks; however, most struggle to perform well in the face of disturbances and uncertain environments. This research proposes a novel approach by integrating the You Only Look Once (YOLO) architecture for object detection with a robust filter for target tracking, addressing issues of disturbances and uncertainties. The YOLO architecture, known for its real-time object detection capabilities, is employed for initial object detection and centroid location. In combination with the detection framework, the sliding innovation filter, a novel robust filter, is implemented and postulated to improve tracking reliability in the face of disturbances. Specifically, the sliding innovation filter is implemented to enhance tracking performance by estimating the optimal centroid location in each frame and updating the object's trajectory. Target tracking traditionally relies on estimation theory techniques like the Kalman filter, and the sliding innovation filter is introduced as a robust alternative particularly suitable for scenarios where a priori information about system dynamics and noise is limited. Experimental simulations in a surveillance scenario demonstrate that the sliding innovation filter-based tracking approach outperforms existing Kalman-based methods, especially in the presence of disturbances. In all, this research contributes a practical and effective approach to object detection and tracking, addressing challenges in real-world, dynamic environments. The comparative analysis with traditional filters provides practical insights, laying the groundwork for future work aimed at advancing multi-object detection and tracking capabilities in diverse applications.

Development of GWIDO: An Augmented Reality-based Mobile Application for Historical Tourism

This research aimed to design and reconstruct a business model for an augmented reality (AR) camera mobile application for historical tourism at Keraton Kasepuhan Cirebon. The goal was to utilize AR technology to provide an immersive and informative experience for tourists. The research addressed several main problems, including navigation and historical information through object tracking, by implementing an online application with features such as Indonesian and English language instructions to better serve domestic and foreign tourists. The research also aimed to investigate the benefits of using AR technology for object tracking and navigation and to explore how these aspects could be related to creating a formula that supports each other in addressing the formulated problems. Through the development of the GWIDO application, a positive impact on the development of historical tourist attractions was observed. This can be seen from the usefulness of its features such as AR navigation, which can be used as a virtual guide. The data collected was used to design and reconstruct the business model, which was implemented and tested to collect additional data for analysis. The final results of the research showed that the AR camera mobile application was effective in providing an immersive and informative experience for tourists. The redesigned business model improved the utilization of AR technology in the tourism industry. Based on the test results, the average response time for object distance between 0.1 meters to 0.5 meters was between 1.45 to 2.07 seconds, and the average time for object distance from visitors was between 3.15 to 4.71 seconds with a confidence level of 95%. Meanwhile, testing for navigation features using augmented reality is very dependent on the internet signal used on the user's device. The level of accuracy of objects that have been placed at certain coordinates is determined by how well the internet network performs, allowing objects to appear precisely according to their coordinates.

Bi-directional Adapter for Multimodal Tracking

Due to the rapid development of computer vision, single-modal (RGB) object tracking has made significant progress in recent years. Considering the limitation of single imaging sensor, multi-modal images (RGB, infrared, etc.) are introduced to compensate for this deficiency for all-weather object tracking in complex environments. However, as acquiring sufficient multi-modal tracking data is hard while the dominant modality changes with the open environment, most existing techniques fail to extract multi-modal complementary information dynamically, yielding unsatisfactory tracking performance. To handle this problem, we propose a novel multi-modal visual prompt tracking model based on a universal bi-directional adapter, cross-prompting multiple modalities mutually. Our model consists of a universal bi-directional adapter and multiple modality-specific transformer encoder branches with sharing parameters. The encoders extract features of each modality separately by using a frozen, pre-trained foundation model. We develop a simple but effective light feature adapter to transfer modality-specific information from one modality to another, performing visual feature prompt fusion in an adaptive manner. With adding fewer (0.32M) trainable parameters, our model achieves superior tracking performance in comparison with both the full fine-tuning methods and the prompt learning-based methods. Our code is available: https://github.com/SparkTempest/BAT.