Research on Video Automatic Feature Extraction Technology Based on Deep Neural Network
Qing Li, Yan Mengqiu, Liu Zhaoping, Lu Jiancheng · 2021 IEEE Asia-Pacific Conference on Image Processing, Electronics and Computers (IPEC) · 2021
With the development of society, people have invested a lot of manpower and material resources in public safety and established various monitoring systems. Video content recognition is to get the theme of video through the analysis of video, and it is an abstract overview of video. Compared with video description and target detection, the results of video content recognition are more concise and abstract, so it can better meet the development needs of information networks with a huge number of videos. Video anomaly detection is to efficiently detect abnormal events from a large number of videos, thus ensuring public safety and preventing dangerous situations. Artificial intelligence system should have the ability to acquire its own knowledge and extract features from the original data. Aiming at the demand of real-time video big data processing ability of video monitoring system, this paper analyzes the automatic video feature extraction technology based on deep neural network, and studies the detection and location of abnormal targets in monitoring video.