Behavior recognition method based on feature mapping and multilayer time interactive attention

Abstract

The invention discloses a behavior recognition method based on feature mapping and multi-layer time interactive attention, and solves the problem that the behavior recognition capability is insufficient due to the fact that time dynamic information modeling is insufficient and the mutual dependency relationship between different frames is ignored in the prior art. The method comprises the following implementation steps: (1) generating a training set; (2) acquiring a depth feature map; (3) constructing a feature mapping matrix; (4) generating a time interaction attention matrix; (5) generating a time interaction attention weighted feature matrix; (6) generating a multi-layer time interactive attention weighted feature matrix; (7) acquiring a feature vector of the video; and (8) carrying out behavior identification on the video. According to the method, the feature mapping matrix is constructed, and multi-layer time interaction attention is provided, so that the accuracy of behavior recognition in the video can be improved.

Description

technical field [0001] The invention belongs to the technical field of video processing, and further relates to a behavior recognition method based on feature mapping and multi-layer temporal interactive attention in the technical field of computer vision. The invention can be used for human body behavior recognition in video. Background technique [0002] Video-based human behavior recognition tasks occupy an important position in the field of computer vision and have broad application prospects. At present, they have been applied in unmanned driving, human-computer interaction, video surveillance and other fields. The goal of human behavior recognition is to judge the category of human behavior in a video, which is essentially a classification problem. In recent years, with the development of deep learning, behavior recognition methods based on deep learning have been widely studied. [0003] South China University of Technology disclosed a human behavior recognition met...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art, and propose a behavior recognition method based on feature mapping and multi-layer temporal interactive attention, which is used to solve the problem of insufficient modeling of temporal dynamic information in the prior art, ignoring different The interdependence between frames leads to the problem of poor behavior recognition ability

Images