Research on Human Action Analysis and Recognition Methods Based on Deep Learning

Abstract

With the vigorous development of the Internet and multimedia technology and the large-scale popularization of video capture devices such as smartphones and surveillance cameras, video data has shown explosive growth. The goal of action recognition is to recognize the action being performed by the human in a video. As a basic but extremely challenging task in computer vision, it has a broad application prospect in many fields, such as human-machine interaction, virtual reality, intelligent video surveillance, and social public security. In this paper, human action analysis and recognition methods based on deep learning are summarized, and several prevailing action recognition algorithms are introduced and categorized in detail. Different from the traditional classification methods, we survey the currently popular algorithms into 2 series from the perspective of feature fusion: 2D convolutional series based action recognition algorithms and 3D convolutional series based action recognition algorithms.