A Optical Flow-Based Fight Behavior Detection Method for Campus Scene

Shu Yang,Yali Li,Shengjin Wang
2023-01-01
Abstract:Campuses contain a large number of facilities that must all be monitored to ensure security. However, most of the existing video surveillance needs to be watched by people, and it is impossible to realize the automatic early warning of some dangerous situations. In this paper, a video-based action detection method is proposed for high-frequency student fight on campus, which uses an optical flow algorithm to perform coarse positioning of the area where fight actions may occur and uses the transformer network to identify the action category of the region of interest. In addition, this paper builds a dataset of fight recognition in middle school campuses for model training, validation and testing. The experimental results show that the method proposed in this paper can locate fight actions relatively accurately and provide real-time early warning.
What problem does this paper attempt to address?