MotionTrack: Learning motion predictor for multiple object tracking.

Significant progress has been achieved in multi-object tracking (MOT) through the evolution of detection and re-identification (ReID) techniques. Despite these advancements, accurately tracking objects in scenarios with homogeneous appearance and heterogeneous motion remains a challenge. This challenge arises from two main factors: the insufficient discriminability of ReID features and the predominant utilization of linear motion models in MOT. In this context, we introduce a novel motion-based tracker, MotionTrack, centered around a learnable motion predictor that relies solely on object trajectory information. This predictor comprehensively integrates two levels of granularity in motion features to enhance the modeling of temporal dynamics and facilitate precise future motion prediction for individual objects. Specifically, the proposed approach adopts a self-attention mechanism to capture token-level information and a Dynamic MLP layer to model channel-level features. MotionTrack is a simple, online tracking approach. Our experimental results demonstrate that MotionTrack yields state-of-the-art performance on datasets such as Dancetrack and SportsMOT, characterized by highly complex object motion.

Query-Based Object Visual Tracking with Parallel Sequence Generation.

Query decoders have been shown to achieve good performance in object detection. However, they suffer from insufficient object tracking performance. Sequence-to-sequence learning in this context has recently been explored, with the idea of describing a target as a sequence of discrete tokens. In this study, we experimentally determine that, with appropriate representation, a parallel approach for predicting a target coordinate sequence with a query decoder can achieve good performance and speed. We propose a concise query-based tracking framework for predicting a target coordinate sequence in a parallel manner, named QPSTrack. A set of queries are designed to be responsible for different coordinates of the tracked target. All the queries jointly represent a target rather than a traditional one-to-one matching pattern between the query and target. Moreover, we adopt an adaptive decoding scheme including a one-layer adaptive decoder and learnable adaptive inputs for the decoder. This decoding scheme assists the queries in decoding the template-guided search features better. Furthermore, we explore the use of the plain ViT-Base, ViT-Large, and lightweight hierarchical LeViT architectures as the encoder backbone, providing a family of three variants in total. All the trackers are found to obtain a good trade-off between speed and performance; for instance, our tracker QPSTrack-B256 with the ViT-Base encoder achieves a 69.1% AUC on the LaSOT benchmark at 104.8 FPS.

MATI: Multimodal Adaptive Tracking Integrator for Robust Visual Object Tracking.

Visual object tracking, pivotal for applications like earth observation and environmental monitoring, encounters challenges under adverse conditions such as low light and complex backgrounds. Traditional tracking technologies often falter, especially when tracking dynamic objects like aircraft amidst rapid movements and environmental disturbances. This study introduces an innovative adaptive multimodal image object-tracking model that harnesses the capabilities of multispectral image sensors, combining infrared and visible light imagery to significantly enhance tracking accuracy and robustness. By employing the advanced vision transformer architecture and integrating token spatial filtering (TSF) and crossmodal compensation (CMC), our model dynamically adjusts to diverse tracking scenarios. Comprehensive experiments conducted on a private dataset and various public datasets demonstrate the model's superior performance under extreme conditions, affirming its adaptability to rapid environmental changes and sensor limitations. This research not only advances visual tracking technology but also offers extensive insights into multisource image fusion and adaptive tracking strategies, establishing a robust foundation for future enhancements in sensor-based tracking systems.

Impact of Perception Errors in Vision-Based Detection and Tracking Pipelines on Pedestrian Trajectory Prediction in Autonomous Driving Systems.

Pedestrian trajectory prediction is crucial for developing collision avoidance algorithms in autonomous driving systems, aiming to predict the future movement of the detected pedestrians based on their past trajectories. The traditional methods for pedestrian trajectory prediction involve a sequence of tasks, including detection and tracking to gather the historical movement of the observed pedestrians. Consequently, the accuracy of trajectory prediction heavily relies on the accuracy of the detection and tracking models, making it susceptible to their performance. The prior research in trajectory prediction has mainly assessed the model performance using public datasets, which often overlook the errors originating from detection and tracking models. This oversight fails to capture the real-world scenario of inevitable detection and tracking inaccuracies. In this study, we investigate the cumulative effect of errors within integrated detection, tracking, and trajectory prediction pipelines. Through empirical analysis, we examine the errors introduced at each stage of the pipeline and assess their collective impact on the trajectory prediction accuracy. We evaluate these models across various custom datasets collected in Taiwan to provide a comprehensive assessment. Our analysis of the results derived from these integrated pipelines illuminates the significant influence of detection and tracking errors on downstream tasks, such as trajectory prediction and distance estimation.

A meta-analysis of performance advantages on athletes in multiple object tracking tasks.

This study compared the multiple object tracking (MOT) performance of athletes vs. non-athletes and expert athletes vs. novice athletes by systematically reviewing and meta-analyzing the literature. A systematic literature search was conducted using five databases for articles published until July 2024. Healthy people were included, specifically classified as athletes and non-athletes, or experts and novices. Potential sources of heterogeneity were selected using a random-effects model. Moderator analyses were also performed. A total of 23 studies were included in this review. Regarding the overall effect, athletes were significantly better at MOT tasks than non-athletes, and experts performed better than novices. Subgroup analyses showed that expert athletes had a significantly larger effect than novices, and that the type of sport significantly moderated the difference in MOT performance between the two groups. Meta-regression revealed that the number of targets and duration of tracking moderated the differences in performance between experts and novices, but did not affect the differences between athletes and non-athletes. This meta-analysis provides evidence of performance advantages for athletes compared with nonathletes, and experts compared with novices in MOT tasks. Moreover, the two effects were moderated by different factors; therefore, future studies should classify participants more specifically according to sports levels.

SiamEFT: adaptive-time feature extraction hybrid network for RGBE multi-domain object tracking.

Integrating RGB and Event (RGBE) multi-domain information obtained by high-dynamic-range and temporal-resolution event cameras has been considered an effective scheme for robust object tracking. However, existing RGBE tracking methods have overlooked the unique spatio-temporal features over different domains, leading to object tracking failure and inefficiency, especally for objects against complex backgrounds. To address this problem, we propose a novel tracker based on adaptive-time feature extraction hybrid networks, namely Siamese Event Frame Tracker (SiamEFT), which focuses on the effective representation and utilization of the diverse spatio-temporal features of RGBE. We first design an adaptive-time attention module to aggregate event data into frames based on adaptive-time weights to enhance information representation. Subsequently, the SiamEF module and cross-network fusion module combining artificial neural networks and spiking neural networks hybrid network are designed to effectively extract and fuse the spatio-temporal features of RGBE. Extensive experiments on two RGBE datasets (VisEvent and COESOT) show that the SiamEFT achieves a success rate of 0.456 and 0.574, outperforming the state-of-the-art competing methods and exhibiting a 2.3-fold enhancement in efficiency. These results validate the superior accuracy and efficiency of SiamEFT in diverse and challenging scenes.

A novel open-access artificial-intelligence-driven platform for CNS drug discovery utilizing adult zebrafish.

BACKGROUND: Although zebrafish are increasingly utilized in biomedicine for CNS disease modelling and drug discovery, this generates big data necessitating objective, precise and reproducible analyses. The artificial intelligence (AI) applications have empowered automated image recognition and video-tracking to ensure more efficient behavioral testing. NEW METHOD: Capitalizing on several AI tools that most recently became available, here we present a novel open-access AI-driven platform to analyze tracks of adult zebrafish collected from in vivo neuropharmacological experiments. For this, we trained the AI system to distinguish zebrafish behavioral patterns following systemic treatment with several well-studied psychoactive drugs - nicotine, caffeine and ethanol. RESULTS: Experiment 1 showed the ability of the AI system to distinguish nicotine and caffeine with 75 % and ethanol with 88 % probability and high (81 %) accuracy following a post-training exposure to these drugs. Experiment 2 further validated our system with additional, previously unexposed compounds (cholinergic arecoline and varenicline, and serotonergic fluoxetine), used as positive and negative controls, respectively. COMPARISON WITH EXISTING METHODS: The present study introduces a novel open-access AI-driven approach to analyze locomotor activity of adult zebrafish. CONCLUSIONS: Taken together, these findings support the value of custom-made AI tools for unlocking full potential of zebrafish CNS drug research by monitoring, processing and interpreting the results of in vivo experiments.

SiamDCFF: Dynamic Cascade Feature Fusion for Vision Tracking.

Establishing an accurate and robust feature fusion mechanism is key to enhancing the tracking performance of single-object trackers based on a Siamese network. However, the output features of the depth-wise cross-correlation feature fusion module in fully convolutional trackers based on Siamese networks cannot establish global dependencies on the feature maps of a search area. This paper proposes a dynamic cascade feature fusion (DCFF) module by introducing a local feature guidance (LFG) module and dynamic attention modules (DAMs) after the depth-wise cross-correlation module to enhance the global dependency modeling capability during the feature fusion process. In this paper, a set of verification experiments is designed to investigate whether establishing global dependencies for the features output by the depth-wise cross-correlation operation can significantly improve the performance of fully convolutional trackers based on a Siamese network, providing experimental support for rational design of the structure of a dynamic cascade feature fusion module. Secondly, we integrate the dynamic cascade feature fusion module into the tracking framework based on a Siamese network, propose SiamDCFF, and evaluate it using public datasets. Compared with the baseline model, SiamDCFF demonstrated significant improvements.

Vehicle Behavior Discovery and Three-Dimensional Object Detection and Tracking Based on Spatio-Temporal Dependency Knowledge and Artificial Fish Swarm Algorithm.

In complex traffic environments, 3D target tracking and detection are often occluded by various stationary and moving objects. When the target is occluded, its apparent characteristics change, resulting in a decrease in the accuracy of tracking and detection. In order to solve this problem, we propose to learn the vehicle behavior from the driving data, predict and calibrate the vehicle trajectory, and finally use the artificial fish swarm algorithm to optimize the tracking results. The experiments show that compared with the CenterTrack method, the proposed method improves the key indicators of MOTA (Multi-Object Tracking Accuracy) in 3D object detection and tracking on the nuScenes dataset, and the frame rate is 26 fps.

Deep Efficient Data Association for Multi-Object Tracking: Augmented with SSIM-Based Ambiguity Elimination.

Recently, to address the multiple object tracking (MOT) problem, we harnessed the power of deep learning-based methods. The tracking-by-detection approach to multiple object tracking (MOT) involves two primary steps: object detection and data association. In the first step, objects of interest are detected in each frame of a video. The second step establishes the correspondence between these detected objects across different frames to track their trajectories. This paper proposes an efficient and unified data association method that utilizes a deep feature association network (deepFAN) to learn the associations. Additionally, the Structural Similarity Index Metric (SSIM) is employed to address uncertainties in the data association, complementing the deep feature association network. These combined association computations effectively link the current detections with the previous tracks, enhancing the overall tracking performance. To evaluate the efficiency of the proposed MOT framework, we conducted a comprehensive analysis of the popular MOT datasets, such as the MOT challenge and UA-DETRAC. The results showed that our technique performed substantially better than the current state-of-the-art methods in terms of standard MOT metrics.

An electric bicycle tracking algorithm for improved traffic management.

This paper proposes an efficient electric bicycle tracking algorithm, EBTrack, utilizing the high-precision and lightweight YOLOv7 as the target detector to enhance the efficiency of illegal detection and recognition of electric bicycles. The EBTrack effectively captures the position and trajectory of electric bicycles in complex traffic monitoring scenarios. Firstly, we introduce the feature extraction network, ResNetEB, specifically designed for feature re-identification of electric bicycles. To maintain real-time performance, feature extraction is performed only when generating new object IDs, minimizing the impact on processing speed. Secondly, for accurate target trajectory prediction, we incorporate an adaptive modulated noise scale Kalman filter. Additionally, considering the uncertainty of electric bicycle entry directions in traffic monitoring scenarios, we design a specialized matching mechanism to reduce frequent ID switching. Finally, to validate the algorithm's effectiveness, we have collected diverse video image data of electric bicycle and urban road traffic in Hefei, Anhui Province, encompassing different perspectives, time periods, and weather conditions. We have trained the proposed detector and have evaluated its tracking performance on this comprehensive dataset. Experimental results demonstrate that EBTrack achieves impressive accuracy, with 89.8 % MOTA (Multiple Object Tracking Accuracy) and 94.2 % IDF1 (ID F1-Score). Furthermore, the algorithm effectively reduces ID switching, significantly improving tracking stability and continuity.

Multiple object tracking training affects the executive function in basketball players: the role of instant feedback.

BACKGROUND: The present study aims to investigate the potential impact of eight sessions of Multiple Object Tracking (MOT) training on the executive function in basketball players. The purpose of the study was primarily to observe the effects of MOT training with and without feedback on the executive function of basketball players. METHODS: A sample of fifty-eight participants was selected from college students enrolled in a university basketball special selection class. The participants were divided into three equal groups. The first group received MOT training with instant feedback and was called feedback group, the second group received MOT training without instant feedback and was called no feedback group, and the third group did not receive any intervention and was called control group. RESULTS: After eight sessions of MOT training, feedback group demonstrated the best performance in the Go/No-go task and the 3-back task. After eight sessions of MOT training, there was no significant difference in test scores on the Stroop task between the feedback and no feedback groups. There was also no significant difference in test scores between the feedback and no feedback groups on the 2-back task after eight sessions of MOT training. The findings of this study suggest that MOT training can effectively enhance the executive function of basketball players. CONCLUSIONS: MOT training was found to enhance the executive function of basketball players, irrespective of whether they received instant feedback. However, the feedback group exhibited superior improvements in the Go/No-go task and the 3-back task.

Reliable object tracking by multimodal hybrid feature extraction and transformer-based fusion.

Visual object tracking, which is primarily based on visible light image sequences, encounters numerous challenges in complicated scenarios, such as low light conditions, high dynamic ranges, and background clutter. To address these challenges, incorporating the advantages of multiple visual modalities is a promising solution for achieving reliable object tracking. However, the existing approaches usually integrate multimodal inputs through adaptive local feature interactions, which cannot leverage the full potential of visual cues, thus resulting in insufficient feature modeling. In this study, we propose a novel multimodal hybrid tracker (MMHT) that utilizes frame-event-based data for reliable single object tracking. The MMHT model employs a hybrid backbone consisting of an artificial neural network (ANN) and a spiking neural network (SNN) to extract dominant features from different visual modalities and then uses a unified encoder to align the features across different domains. Moreover, we propose an enhanced transformer-based module to fuse multimodal features using attention mechanisms. With these methods, the MMHT model can effectively construct a multiscale and multidimensional visual feature space and achieve discriminative feature modeling. Extensive experiments demonstrate that the MMHT model exhibits competitive performance in comparison with that of other state-of-the-art methods. Overall, our results highlight the effectiveness of the MMHT model in terms of addressing the challenges faced in visual object tracking tasks.

Enhancing Badminton Game Analysis: An Approach to Shot Refinement via a Fusion of Shuttlecock Tracking and Hit Detection from Monocular Camera.

Extracting the flight trajectory of the shuttlecock in a single turn in badminton games is important for automated sports analytics. This study proposes a novel method to extract shots in badminton games from a monocular camera. First, TrackNet, a deep neural network designed for tracking small objects, is used to extract the flight trajectory of the shuttlecock. Second, the YOLOv7 model is used to identify whether the player is swinging. As both TrackNet and YOLOv7 may have detection misses and false detections, this study proposes a shot refinement algorithm to obtain the correct hitting moment. By doing so, we can extract shots in rallies and classify the type of shots. Our proposed method achieves an accuracy of 89.7%, a recall rate of 91.3%, and an F1 rate of 90.5% in 69 matches, with 1582 rallies of the Badminton World Federation (BWF) match videos. This is a significant improvement compared to the use of TrackNet alone, which yields 58.8% accuracy, 93.6% recall, and 72.3% F1 score. Furthermore, the accuracy of shot type classification at three different thresholds is 72.1%, 65.4%, and 54.1%. These results are superior to those of TrackNet, demonstrating that our method effectively recognizes different shot types. The experimental results demonstrate the feasibility and validity of the proposed method.

DOT-SLAM: A Stereo Visual Simultaneous Localization and Mapping (SLAM) System with Dynamic Object Tracking Based on Graph Optimization.

Most visual simultaneous localization and mapping (SLAM) systems are based on the assumption of a static environment in autonomous vehicles. However, when dynamic objects, particularly vehicles, occupy a large portion of the image, the localization accuracy of the system decreases significantly. To mitigate this challenge, this paper unveils DOT-SLAM, a novel stereo visual SLAM system that integrates dynamic object tracking through graph optimization. By integrating dynamic object pose estimation into the SLAM system, the system can effectively utilize both foreground and background points for ego vehicle localization and obtain a static feature points map. To rectify the inaccuracies in depth estimation from stereo disparity directly on the foreground points of dynamic objects due to their self-similarity characteristics, a coarse-to-fine depth estimation method based on camera-road plane geometry is presented. This method uses rough depth to guide fine stereo matching, thereby obtaining the 3 dimensions (3D)spatial positions of feature points on dynamic objects. Subsequently, by establishing constraints on the dynamic object's pose using the road plane and non-holonomic constraints (NHCs) of the vehicle, reducing the initial pose uncertainty of dynamic objects leads to more accurate dynamic object initialization. Finally, by considering foreground points, background points, the local road plane, the ego vehicle pose, and dynamic object poses as optimization nodes, through the establishment and joint optimization of a nonlinear model based on graph optimization, accurate six degrees of freedom (DoFs) pose estimations are obtained for both the ego vehicle and dynamic objects. Experimental validation on the KITTI-360 dataset demonstrates that DOT-SLAM effectively utilizes features from the background and dynamic objects in the environment, resulting in more accurate vehicle trajectory estimation and a static environment map. Results obtained from a real-world dataset test reinforce the effectiveness.

Improved DeepSORT-Based Object Tracking in Foggy Weather for AVs Using Sematic Labels and Fused Appearance Feature Network.

The presence of fog in the background can prevent small and distant objects from being detected, let alone tracked. Under safety-critical conditions, multi-object tracking models require faster tracking speed while maintaining high object-tracking accuracy. The original DeepSORT algorithm used YOLOv4 for the detection phase and a simple neural network for the deep appearance descriptor. Consequently, the feature map generated loses relevant details about the track being matched with a given detection in fog. Targets with a high degree of appearance similarity on the detection frame are more likely to be mismatched, resulting in identity switches or track failures in heavy fog. We propose an improved multi-object tracking model based on the DeepSORT algorithm to improve tracking accuracy and speed under foggy weather conditions. First, we employed our camera-radar fusion network (CR-YOLOnet) in the detection phase for faster and more accurate object detection. We proposed an appearance feature network to replace the basic convolutional neural network. We incorporated GhostNet to take the place of the traditional convolutional layers to generate more features and reduce computational complexities and costs. We adopted a segmentation module and fed the semantic labels of the corresponding input frame to add rich semantic information to the low-level appearance feature maps. Our proposed method outperformed YOLOv5 + DeepSORT with a 35.15% increase in multi-object tracking accuracy, a 32.65% increase in multi-object tracking precision, a speed increase by 37.56%, and identity switches decreased by 46.81%.

Zebrafishtracker3D: A 3D skeleton tracking algorithm for multiple zebrafish based on particle matching.

Zebrafish are considered as model organisms in biological and medical research because of their high degree of homology with human genes. Automatic behavioral analysis of multiple zebrafish based on visual tracking is expected to improve research efficiency. However, vision-based multi-object tracking algorithms often suffer from data loss owing to mutual occlusion. In addition, simply tracking zebrafish as points is not sufficient-more detailed information, which is required for research on zebrafish behavior. In this paper, we propose Zebrafishtracker3D, which utilizes a skeleton stability strategy to reduce detection error caused by frequent overlapping of multiple zebrafish effectively and estimates zebrafish skeletons using head coordinates in the top view. Further, we transform the front- and top-view matching task into an optimization problem and propose a particle-matching method to perform 3D tracking. The robustness of the algorithm with respect to occlusion is estimated on the dataset comprising two and three zebrafish. Experimental results demonstrate that the proposed algorithm exhibits a multiple object tracking accuracy (MOTA) exceeding 90% in the top view and a 3D tracking matching accuracy exceeding 90% in the complex videos with frequent overlapping. It is noteworthy that each instance in the trace saves its skeleton. In addition, Zebrafishtracker3D is applied in the zebrafish courtship experiment, establishes the stability of the method in applications of life science, and proves that the data can be used for behavioral analysis. Zebrafishtracker3D is the first algorithm that realizes 3D skeleton tracking of multiple zebrafish simultaneously.

Enhancing visual communication through representation learning.

Introduction: This research aims to address the challenges in model construction for the Extended Mind for the Design of the Human Environment. Specifically, we employ the ResNet-50, LSTM, and Object Tracking Algorithms approaches to achieve collaborative construction of high-quality virtual assets, image optimization, and intelligent agents, providing users with a virtual universe experience in the context of visual communication. Methods: Firstly, we utilize ResNet-50 as a convolutional neural network model for generating virtual assets, including objects, characters, and environments. By training and fine-tuning ResNet-50, we can generate virtual elements with high realism and rich diversity. Next, we use LSTM (Long Short-Term Memory) for image processing and analysis of the generated virtual assets. LSTM can capture contextual information in image sequences and extract/improve the details and appearance of the images. By applying LSTM, we further enhance the quality and realism of the generated virtual assets. Finally, we adopt Object Tracking Algorithms to track and analyze the movement and behavior of virtual entities within the virtual environment. Object Tracking Algorithms enable us to accurately track the positions and trajectories of objects, characters, and other elements, allowing for realistic interactions and dynamic responses. Results and discussion: By integrating the technologies of ResNet-50, LSTM, and Object Tracking Algorithms, we can generate realistic virtual assets, optimize image details, track and analyze virtual entities, and train intelligent agents, providing users with a more immersive and interactive visual communication-driven metaverse experience. These innovative solutions have important applications in the Extended Mind for the Design of the Human Environment, enabling the creation of more realistic and interactive virtual worlds.

Effects of 6-Week Motor-Cognitive Agility Training on Football Test Performance in Adult Amateur Players - A Three-Armed Randomized Controlled Trial.

Agility, defined as the ability to rapidly respond to unforeseen events, constitutes a central performance component in football. Existing agility training approaches often focus on change of direction that does not reflect the complex motor-cognitive demands on the pitch. The objective of this study is to examine the effects of a novel motor-cognitive dual-task agility training (Multiple-object tracking integrated into agility training) on agility and football-specific test performance parameters, compared to agility and a change of direction (COD) training. Adult male amateur football players (n = 42; age: 27±6; height: 181±7cm; weight: 80±12kg) were randomly allocated to one of the three intervention groups (COD, agility, agility + multiple object tracking). The Loughborough Soccer Passing Test (LSPT), a dribbling test with/without cognitive task as well as the Random Star Run (with/without ball) and the modified T-Test were assessed before and after a 6-week training period. Time effects within the T-Test (F = 83.9; p < 0.001; η2 = 0.68) and dribbling test without cognitive task (F = 23.9; p < 0.001; η2 = 0.38) with improvements of all intervention groups (p < 0.05) were found. Dribbling with cognitive task revealed a time effect (F = 7.8; p = 0.008; η2 = 0.17), with improvements exclusively in the agility and dual-task agility groups (p < 0.05). Random Star Run with and without ball exhibited a time (F = 38.8; p < 0.001; η2 = 0.5; F = 82.7; p < 0.001; η2 = 0.68) and interaction effect (F = 14.14; p < 0.001; η2 = 0.42; F = 27.8; p < 0.001; η2 = 0.59), with improvements for the agility and dual-task agility groups. LSPT showed no time, group or interaction effect. The effects of change of direction training are limited to change of direction and dribbling test performance within preplanned scenarios. In contrast, motor-cognitive agility interventions result in notable enhancements in football-specific and agility tests, incorporating decision-making and multitasking components. No differences were observed between agility and agility + multiple object tracking. To achieve a transfer to game-relevant performance, coaches should focus on integrating cognitive challenges into motor training.

An Improved Two-Shot Tracking Algorithm for Dynamics Analysis of Natural Killer Cells in Tumor Contexts.

Natural killer cells (NKCs) are non-specific immune lymphocytes with diverse morphologies. Their broad killing effect on cancer cells has led to increased attention towards activating NKCs for anticancer immunotherapy. Consequently, understanding the motion characteristics of NKCs under different morphologies and modeling their collective dynamics under cancer cells has become crucial. However, tracking small NKCs in complex backgrounds poses significant challenges, and conventional industrial tracking algorithms often perform poorly on NKC tracking datasets. There remains a scarcity of research on NKC dynamics. In this paper, we utilize deep learning techniques to analyze the morphology of NKCs and their key points. After analyzing the shortcomings of common industrial multi-object tracking algorithms like DeepSORT in tracking natural killer cells, we propose Distance Cascade Matching and the Re-Search method to improve upon existing algorithms, yielding promising results. Through processing and tracking over 5000 frames of images, encompassing approximately 300,000 cells, we preliminarily explore the impact of NKCs' cell morphology, temperature, and cancer cell environment on NKCs' motion, along with conducting basic modeling. The main conclusions of this study are as follows: polarized cells are more likely to move along their polarization direction and exhibit stronger activity, and the maintenance of polarization makes them more likely to approach cancer cells; under equilibrium, NK cells display a Boltzmann distribution on the cancer cell surface.