A Review of Human Behavior Recognition Based on Deep Learning

Abstract

Human behavior recognition is a research hotspot in the field of computer vision. This paper introduces and analyzes the methods of human motion and behavior recognition in the field of video in recent years, describes the main ways to build the network and the methods to optimize the network parameters to improve the model recognition effect, as well as the commonly used data sets. At present, human behavior recognition still needs to be further studied, and there are still great challenges in accuracy, number of model parameters and calculation amount.