Real-Time Human Action Recognition Using Deep Learning

Abstract

The future of computer vision lies in deep learning to develop machines to solve our human problems. One of the most important areas of research is smart video surveillance. This feature is related to the study and recognition of movements, and it's used in many fields, like security, sports, medicine, and a whole lot of new applications. The study and analysis of human activity is very important to improve because it is a very sensitive field, like in security, the human needs a machine's help a lot; and in recent years, developers have adopted many advanced algorithms to discover the type of movements humans preform, and the results differ from one to another. The most important part of human activity recognition is real time, so one can detect any issue, like a medical problem, in time. In this regard, the authors will use methods of deep learning to reach a good result of recognition of the nature of human action in real time clips.