Human Action Recognition Based on Model Structure

Abstract

Currently, Human Activity Recognition is a research hotspot in the field of machine vision, it involves knowledge of image processing, pattern recognition, artificial intelligence and many other disciplines. Video-based Human Activity Recognition including human area detection, movement and gesture segmentation, objective analysis and behavior understand for activity recognition and so on. In this paper, according to the feature of running, jogging and walking, we structure the model of these actions, then we will compare with the samples and model to distinguish the actions. At first, we will have some process for the images we get from the video. Then, we will find out the features of each action and extract these features. Finally, we will build the models of running and jogging on the basis of these features.