The design of an intelligent monitoring system for human hand behaviors

Intelligent monitoring to observe human behaviors and report anomaly activities is a common application of computer vision technologies. However, to the authors knowledge, there has not been a widely accepted structure of building such a system with the fast-developing deep learning method. Within the author's knowledge, current works focus on industrial and traffic conditions, such as measuring the speeds of vehicles in highways, or to help arrange agriculture productions. This paper presents an efficient approach to applying deep learning techniques in such systems to help analyze human behaviors. Specifically, we combined the advancing object detection, pose estimation and image classification methods in our work to recognize some anomaly or special behaviors in determined scenes.