SORT-YM: An Algorithm of Multi-Object Tracking with YOLOv4-Tiny and Motion Prediction

Multi-object tracking (MOT) is a significant and widespread research field in image processing and computer vision. The goal of the MOT task consists in predicting the complete tracklets of multiple objects in a video sequence. There are usually many challenges that degrade the performance of the algorithm in the tracking process, such as occlusion and similar objects. However, the existing MOT algorithms based on the tracking-by-detection paradigm struggle to accurately predict the location of the objects that they fail to track in complex scenes, leading to tracking performance decay, such as an increase in the number of ID switches and tracking drifts. To tackle those difficulties, in this study, we design a motion prediction strategy for predicting the location of occluded objects. Since the occluded objects may be legible in earlier frames, we utilize the speed and location of the objects in the past frames to predict the possible location of the occluded objects. In addition, to improve the tracking speed and further enhance the tracking robustness, we utilize efficient YOLOv4-tiny to produce the detections in the proposed algorithm. By using YOLOv4-tiny, the tracking speed of our proposed method improved significantly. The experimental results on two widely used public datasets show that our proposed approach has obvious advantages in tracking accuracy and speed compared with other comparison algorithms. Compared to the Deep SORT baseline, our proposed method has a significant improvement in tracking performance.

Download Full-text

Robust Object Tracking Algorithm for Autonomous Vehicles in Complex Scenes

Remote Sensing ◽

10.3390/rs13163234 ◽

2021 ◽

Vol 13 (16) ◽

pp. 3234

Author(s):

Jingwei Cao ◽

Chuanxue Song ◽

Shixin Song ◽

Feng Xiao ◽

Xu Zhang ◽

...

Keyword(s):

Object Tracking ◽

Autonomous Vehicles ◽

Feature Fusion ◽

Environmental Perception ◽

Tracking Algorithm ◽

Tracking Performance ◽

Tracking Accuracy ◽

Complex Scenes ◽

Driving Assistance ◽

Public Datasets

Object tracking is an essential aspect of environmental perception technology for autonomous vehicles. The existing object tracking algorithms can only be applied well to simple scenes. When the scenes become complex, the algorithms have poor tracking performance and insufficient robustness, and the problems of tracking drift and object loss are prone to occur. Therefore, a robust object tracking algorithm for autonomous vehicles in complex scenes is proposed. Firstly, we study the Siam-FC network and related algorithms, and analyze the problems that need to be addressed in object tracking. Secondly, the construction of a double-template Siamese network model based on multi-feature fusion is described, as is the use of the improved MobileNet V2 as the feature extraction backbone network, and the attention mechanism and template online update mechanism are introduced. Finally, relevant experiments were carried out based on public datasets and actual driving videos, with the aim of fully testing the tracking performance of the proposed algorithm on different objects in a variety of complex scenes. The results showed that, compared with other algorithms, the proposed algorithm had high tracking accuracy and speed, demonstrated stronger robustness and anti-interference abilities, and could still accurately track the object in real time without the introduction of complex structures. This algorithm can be effectively applied in intelligent vehicle driving assistance, and it will help to promote the further development and improvement of computer vision technology in the field of environmental perception.

Download Full-text

Multiple objects tracking in the UAV system based on hierarchical deep high-resolution network

Multimedia Tools and Applications ◽

10.1007/s11042-020-10427-1 ◽

2021 ◽

Author(s):

Wei Huang ◽

Xiaoshu Zhou ◽

Mingchao Dong ◽

Huaiyu Xu

Keyword(s):

High Resolution ◽

Object Tracking ◽

High Performance ◽

State Of The Art ◽

Class Imbalance ◽

Unified Framework ◽

Multiple Objects ◽

Tracking Process ◽

Objects Tracking ◽

Different Types

AbstractRobust and high-performance visual multi-object tracking is a big challenge in computer vision, especially in a drone scenario. In this paper, an online Multi-Object Tracking (MOT) approach in the UAV system is proposed to handle small target detections and class imbalance challenges, which integrates the merits of deep high-resolution representation network and data association method in a unified framework. Specifically, while applying tracking-by-detection architecture to our tracking framework, a Hierarchical Deep High-resolution network (HDHNet) is proposed, which encourages the model to handle different types and scales of targets, and extract more effective and comprehensive features during online learning. After that, the extracted features are fed into different prediction networks for interesting targets recognition. Besides, an adjustable fusion loss function is proposed by combining focal loss and GIoU loss to solve the problems of class imbalance and hard samples. During the tracking process, these detection results are applied to an improved DeepSORT MOT algorithm in each frame, which is available to make full use of the target appearance features to match one by one on a practical basis. The experimental results on the VisDrone2019 MOT benchmark show that the proposed UAV MOT system achieves the highest accuracy and the best robustness compared with state-of-the-art methods.

Download Full-text

CenterTrack3D: Improved CenterTrack more Suitable for 3D Objects

Journal of Autonomous Vehicles and Systems ◽

10.1115/1.4050863 ◽

2021 ◽

pp. 1-32

Author(s):

Lipeng Gu ◽

Shaoyuan Sun ◽

Xunhua Liu ◽

Xiang Li

Keyword(s):

Object Tracking ◽

Research Field ◽

Unmanned Vehicles ◽

Current Frame ◽

3D Objects ◽

Detection And Tracking ◽

Tracking Algorithms ◽

Occluded Objects ◽

Previous Frame ◽

3D Information

Abstract Compared with 2D multi-object tracking algorithms, 3D multi-object tracking algorithms have more research significance and broad application prospects in the unmanned vehicles research field. Aiming at the problem of 3D multi-object detection and tracking, in this paper, the multi-object tracker CenterTrack, which focuses on 2D multi-object tracking task while ignoring object 3D information, is improved mainly from two aspects of detection and tracking, and the improved network is called CenterTrack3D. In terms of detection, CenterTrack3D uses the idea of attention mechanism to optimize the way that the previous-frame image and the heatmap of previous-frame tracklets are added to the current-frame image as input, and second convolutional layer of the output head is replaced by dynamic convolution layer, which further improves the ability to detect occluded objects. In terms of tracking, a cascaded data association algorithm based on 3D Kalman filter is proposed to make full use of the 3D information of objects in the image and increase the robustness of the 3D multi-object tracker. The experimental results show that, compared with the original CenterTrack and the existing 3D multi-object tracking methods, CenterTrack3D achieves 88.75% MOTA for cars and 59.40% MOTA for pedestrians and is very competitive on the KITTI tracking benchmark test set.

Download Full-text

Least Trimmed Squares Approach to Lucas-Kanade Algorithm in Object Tracking Problems

Mathematical Problems in Engineering ◽

10.1155/2013/324824 ◽

2013 ◽

Vol 2013 ◽

pp. 1-6

Author(s):

Yih-Lon Lin

Keyword(s):

Cost Function ◽

Object Tracking ◽

Vehicle Tracking ◽

Tracking Accuracy ◽

Least Trimmed Squares ◽

Environment Adaptation ◽

Tracking Process ◽

Important Research Topic ◽

Process Simulations ◽

The Cost

The object tracking problem is an important research topic in computer vision. For real applications such as vehicle tracking and face tracking, there are many efficient and real-time algorithms. In this study, we will focus on the Lucas-Kanade (LK) algorithm for object tracking. Although this method is time consuming, it is effective in tracking accuracy and environment adaptation. In the standard LK method, the sum of squared errors is used as the cost function, while least trimmed squares is adopted as the cost function in this study. The resulting estimator is robust against outliers caused by noises and occlusions in the tracking process. Simulations are provided to show that the proposed algorithm outperforms the standard LK method in the sense that it is robust against the outliers in the object tracking problems.

Download Full-text

The Application of Image Processing to Solve Occlusion Issue in Object Tracking

MATEC Web of Conferences ◽

10.1051/matecconf/201815203001 ◽

2018 ◽

Vol 152 ◽

pp. 03001

Author(s):

Yun Zhe Cheong ◽

Wei Jen Chew

Keyword(s):

Object Tracking ◽

Moving Objects ◽

Object Segmentation ◽

Target Object ◽

Occlusion Handling ◽

Multiple Objects ◽

The Road ◽

Occluded Objects ◽

On The Road ◽

Tracking Object

Object tracking is a computer vision field that involves identifying and tracking either a single or multiple objects in an environment. This is extremely useful to help observe the movements of the target object like people in the street or cars on the road. However, a common issue with tracking an object in an environment with many moving objects is occlusion. Occlusion can cause the system to lose track of the object being tracked or after overlapping, the wrong object will be tracked instead. In this paper, a system that is able to correctly track occluded objects is proposed. This system includes algorithms such as foreground object segmentation, colour tracking, object specification and occlusion handling. An input video is input to the system and every single frame of the video is analysed. The foreground objects are segmented with object segmentation algorithm and tracked with the colour tracking algorithm. An ID is assigned to each tracked object. Results obtained shows that the proposed system is able to continuously track an object and maintain the correct identity even after is has been occluded by another object.

Download Full-text

Parallel Correlation Filters for Real-Time Visual Tracking

Sensors ◽

10.3390/s19102362 ◽

2019 ◽

Vol 19 (10) ◽

pp. 2362 ◽

Cited By ~ 5

Author(s):

Yijin Yang ◽

Yihong Zhang ◽

Demin Li ◽

Zhijie Wang

Keyword(s):

Object Tracking ◽

Real Time ◽

Research Field ◽

Tracking Performance ◽

Correlation Filter ◽

Visual Object ◽

Correlation Filters ◽

Illumination Variation ◽

Visual Object Tracking ◽

Appearance Changes

Correlation filter-based methods have recently performed remarkably well in terms of accuracy and speed in the visual object tracking research field. However, most existing correlation filter-based methods are not robust to significant appearance changes in the target, especially when the target undergoes deformation, illumination variation, and rotation. In this paper, a novel parallel correlation filters (PCF) framework is proposed for real-time visual object tracking. Firstly, the proposed method constructs two parallel correlation filters, one for tracking the appearance changes in the target, and the other for tracking the translation of the target. Secondly, through weighted merging the response maps of these two parallel correlation filters, the proposed method accurately locates the center position of the target. Finally, in the training stage, a new reasonable distribution of the correlation output is proposed to replace the original Gaussian distribution to train more accurate correlation filters, which can prevent the model from drifting to achieve excellent tracking performance. The extensive qualitative and quantitative experiments on the common object tracking benchmarks OTB-2013 and OTB-2015 have demonstrated that the proposed PCF tracker outperforms most of the state-of-the-art trackers and achieves a high real-time tracking performance.

Download Full-text

Multiple Object Tracking using YOLO-based Detector

Journal of Imaging Science and Technology ◽

10.2352/j.imagingsci.technol.2021.65.4.040401 ◽

2021 ◽

Author(s):

Shinfeng D. Lin ◽

Tingyu Chang ◽

Wensheng Chen

Keyword(s):

Object Tracking ◽

State Of The Art ◽

Multiple Object Tracking ◽

Motion Prediction ◽

Multiple Object ◽

Bounding Box ◽

Video Frames ◽

Tracking Process ◽

Tracking By Detection ◽

Bounding Boxes

In computer vision, multiple object tracking (MOT) plays a crucial role in solving many important issues. A common approach of MOT is tracking by detection. Tracking by detection includes occlusions, motion prediction, and object re-identification. From the video frames, a set of detections is extracted for leading the tracking process. These detections are usually associated together for assigning the same identifications to bounding boxes holding the same target. In this article, MOT using YOLO-based detector is proposed. The authors’ method includes object detection, bounding box regression, and bounding box association. First, the YOLOv3 is exploited to be an object detector. The bounding box regression and association is then utilized to forecast the object’s position. To justify their method, two open object tracking benchmarks, 2D MOT2015 and MOT16, were used. Experimental results demonstrate that our method is comparable to several state-of-the-art tracking methods, especially in the impressive results of MOT accuracy and correctly identified detections.

Download Full-text

Multi-Object Tracking and Occlusion Reasoning Based on Adaptive Weighing Particle Filter

New Mathematics and Natural Computation ◽

10.1142/s1793005714500136 ◽

2014 ◽

Vol 10 (03) ◽

pp. 225-238 ◽

Cited By ~ 2

Author(s):

Chenguang Liu ◽

Heng-Da Cheng

Keyword(s):

Particle Filter ◽

Object Tracking ◽

Challenging Problem ◽

Multiple Objects ◽

Occluded Objects ◽

Occlusion Model ◽

Occlusion Reasoning

Multi-object tracking has significant merit to our society. However, interactions among objects result in complex spatial occlusions, which gives rise to a challenging problem in tracking. We propose an adaptive weighing particle filter (AWPT) for tracking multiple objects and reasoning the occlusions among them. A weighing-occlusion modeling-weighing procedure is developed to adaptively weigh the particles. Moreover, we propose a stack occlusion model and define the operations on it to maintain the occlusion relationship. The experiments exhibit that the proposed method can effectively track fully occluded objects and reason about the occlusion relationships among them.

Download Full-text

HOTA: A Higher Order Metric for Evaluating Multi-object Tracking

International Journal of Computer Vision ◽

10.1007/s11263-020-01375-2 ◽

2020 ◽

Author(s):

Jonathon Luiten ◽

Aljos̆a Os̆ep ◽

Patrick Dendorfer ◽

Philip Torr ◽

Andreas Geiger ◽

...

Keyword(s):

Object Tracking ◽

Higher Order ◽

Tracking Performance ◽

Tracking Accuracy ◽

Basic Error ◽

Visual Evaluation ◽

Order Tracking ◽

Accurate Detection ◽

Evaluation Metric ◽

Error Types

Abstract Multi-object tracking (MOT) has been notoriously difficult to evaluate. Previous metrics overemphasize the importance of either detection or association. To address this, we present a novel MOT evaluation metric, higher order tracking accuracy (HOTA), which explicitly balances the effect of performing accurate detection, association and localization into a single unified metric for comparing trackers. HOTA decomposes into a family of sub-metrics which are able to evaluate each of five basic error types separately, which enables clear analysis of tracking performance. We evaluate the effectiveness of HOTA on the MOTChallenge benchmark, and show that it is able to capture important aspects of MOT performance not previously taken into account by established metrics. Furthermore, we show HOTA scores better align with human visual evaluation of tracking performance.

Download Full-text

Siamese Tracking from Single Point Initialization

Sensors ◽

10.3390/s19030514 ◽

2019 ◽

Vol 19 (3) ◽

pp. 514 ◽

Cited By ~ 1

Author(s):

Zheng Xu ◽

Haibo Luo ◽

Bin Hui ◽

Zheng Chang

Keyword(s):

Object Tracking ◽

State Of The Art ◽

Single Point ◽

Contour Detection ◽

Tracking Accuracy ◽

Object Contour ◽

Start Up ◽

Tracking Process ◽

Real Time Tracking ◽

Target Locating

Recently, we have been concerned with locating and tracking vehicles in aerial videos. Vehicles in aerial videos usually have small sizes due to use of cameras from a remote distance. However, most of the current methods use a fixed bounding box region as the input of tracking. For the purpose of target locating and tracking in our system, detecting the contour of the target is utilized and can help with improving the accuracy of target tracking, because a shape-adaptive template segmented by object contour contains the most useful information and the least background for object tracking. In this paper, we propose a new start-up of tracking by clicking on the target, and implement the whole tracking process by modifying and combining a contour detection network and a fully convolutional Siamese tracking network. The experimental results show that our algorithm has significantly improved tracking accuracy compared to the state-of-the-art regarding vehicle images in both OTB100 and DARPA datasets. We propose utilizing our method in real time tracking and guidance systems.

Download Full-text