Human behavior prediction method and system based on adaptive graph convolutional adversarial network

Abstract

The present disclosure proposes a method and system for predicting human behavior based on an adaptive graph convolutional adversarial network. The method includes the following steps: acquiring a human behavior action sequence and segmenting it according to different observation ratios; The AGCN-AL network and the local network respectively perform behavior prediction; the prediction results of the local network and the AGCN-AL network are fused as the final behavior prediction result; the AGCN-AL network includes a graph adaptive graph convolutional network module. The feature extraction network of , and the discriminator and the classifier respectively connected with the feature extraction network, and the local network includes the feature extraction network and the classifier connected in sequence. The present disclosure integrates the adaptive graph convolution confrontation network AGCN-AL and the local network of the adaptive graph convolution confrontation network AGCN-AL for human skeleton data, realizes the fusion of prediction results of global information and local information, and improves the performance of human behavior prediction. accuracy.

Description

technical field [0001] The present disclosure relates to the technical field of image processing, and in particular, to a method and system for predicting human behavior based on an adaptive graph convolutional confrontation network. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] The purpose of behavior prediction is to identify an action before it ends. Behavior prediction has a wide range of application prospects in actual production and life. For example, in monitoring scenarios such as security monitoring and medical monitoring, it can detect and prevent the occurrence of dangerous events as early as possible, and improve the intelligence level of monitoring. Behavior prediction can also be applied to assisted driving, autonomous driving, human-machine collaboration and other fields. It can not only provide early warning of dangerou...

AI Technical Summary

Problems solved by technology

[0005] The inventor found that there are few related algorithms for behavior prediction based on human skeleton data, and the existing methods of behavior prediction based on skeleton data either have poor prediction effect or complex network structure, which is not conducive to practical application

Images