Human body behavior recognition method based on interlaced attention enhancing network

Abstract

The invention provides a human body behavior recognition method based on an interlaced attention enhancing network, and solves the problems that local information is ignored in the prior art and is easily interfered by a large amount of redundant background information and information irrelevant to behaviors in a video, and the behavior recognition capability is insufficient. The method comprisesthe following steps: (1) generating a training set; (2) obtaining a low-level feature map and a high-level feature map; (3) constructing a hierarchical complementary attention module; (4) constructinga local attention enhancing module; (5) establishing a classification network; (6) constructing an interlaced attention enhancing network; (7) constructing a loss function of the interlaced attentionenhancing network; (8) training an interlaced attention enhancing network; and (9) identifying behaviors in the video images. According to the method, the interlaced attention enhancing network and the loss function thereof are constructed, so that the accuracy of behavior recognition can be improved.

Description

technical field [0001] The invention belongs to the technical field of video processing, and further relates to a human behavior recognition method based on an interlaced enhanced attention network in the technical field of computer vision. The invention can be used to identify the behavior category of human body from video. Background technique [0002] In recent years, with the development of artificial intelligence and computer vision, video-based human behavior recognition has been widely used in intelligent video surveillance, human-computer interaction, unmanned driving and other technical fields. The main goal of human behavior recognition is to judge the category of human behavior in a video. Therefore, human behavior recognition can also be regarded as a classification problem in which the input is video and the output is behavior category. At present, convolutional neural network has become the mainstream method in human action recognition due to its powerful ima...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the deficiencies in the above-mentioned prior art, and propose a human behavior recognition method based on interlaced enhanced attention network, which is used to solve the problem that the prior art ignores the local detail information in the video frame, which is easy to be detected in the video. There is a large amount of redundant background information and information interference irrelevant to behavior, which leads to the problem of insufficient recognition ability of behavior

Images