Survey of video object detection algorithms based on deep learning
Zheng Lu, Tongtong Zhou, Rongqi Jiang, Yueping Peng · 2021 4th International Conference on Algorithms, Computing and Artificial Intelligence · 2021
Video object detection is based on the static image detection, and combines the time dimension characteristics to further improve the performance of object detection. As an extension of the field of object detection, video object detection has gradually become more and more popular in recent years, and the video object detection algorithm based on deep learning is expanding. In this paper, the main video object detection algorithms are summarized. Firstly, based on the relevant principles of the algorithm, the main algorithms' motivation, core ideas, network architecture and innovation points are classified. Based on the main structure, design ideas and implementation effect, the advantages and disadvantages of the algorithms are compared. Based on the detection accuracy of the algorithm on ImageNet VID data set, the performance of the algorithms are analyzed. Finally, the problems and challenges in the field of video object detection are proposed, and research direction in the future is prospected.