An Overview of Object Detection and Tracking Algorithms

Engineering Proceedings

article

Kehao Du¹  Alexander Bobkov¹ 

[1] Department of Informatics and Control Systems, Bauman Moscow State Technical University

关键词: region proposal;    R-CNN;    Fast R-CNN;    Faster R-CNN;    YOLO;    MOSSE;    KCF;    SORT;   

DOI  :  10.3390/engproc2023033022

来源: mdpi

【 摘 要 】

With the development of information technology, the vision-based detection and tracking of moving objects is gradually penetrating into all aspects of people’s lives, and its importance is becoming more prominent, attracting more and more scientists and research institutions at home and abroad to participate in research in this field. With in-depth research into vision-based object detection and tracking, various superior algorithms have appeared in recent years. In this article, we attempt to compare some of the classic algorithms in this area of detection and tracking that have appeared recently. This article examines and summarizes two areas: the detection and the tracking of moving objects. First, we divide object detection into one-stage algorithms and two-stage algorithms depending on whether a region proposal should be generated, and we accordingly outline some commonly used object detection algorithms. Second, we separate object tracking into the KCF and SORT algorithms according to the differences in the underlying algorithms.

【 授权许可】

Unknown