Development of HAR using Spatio Temporal Machine Learning Techniques

Action recognition of human aid us to spot their action their nature, and emotional state, that is hard for extraction. Hence, it plays a decisive role in person’s interaction and communal relations. One of the primary subjects of study in the scientific fields of computer vision and machine learning is the human ability to recognize another person’s activity. By the results of this analysis, several activity recognition system is needed for several applications, together with video surveillance systems, human-computer interaction, and robotics for human behavior characterization. For the analysis of image and video, human activity recognition (HAR) is an important analysis direction. In this paper, we offer an embrace survey of the recent development techniques, including strategies, systems, and quantitative analysis of the performance of HAR. By the experimental results it shows that our technique will remarkably improve classification, interpretation, and retrieval performance for the video pictures. The novelty of this paper is bipartite. Firstly, the video pictures of human is captured. Secondly, different types of actions performed by humans are going to be identified.