The Human Body Detection In A Number Of Key Technology Research

Object detection is one of the hottest topics in computer vision areas. It is an obvious denotation for computer intelligence as well. This dissertation focuses the research on human detection, which is an extra example of object detection. Human detection is the key point for automatic recognition of human activities and image understanding. It has a wide use in survenillance system and human computer interaction.Human body is non-rigid body. It is challenging due to the appearance variation (color, texture), the presence of pose ambiguity, and the different views of human body in an image. Therefore it is arguably the most difficulty problem in computer vision. This dissertation studies the cutting edge methods for human detection. Compared with our research, the probability optimization methods are explored and improved, which is then applied to the human detection problem in dissertation. The results show that this method is promising. The main works of this dissertation are as follow:1. The basic methods of object detection are studied. Especially feature extraction and statistical learning are included. The key problem is also explored.2. The state-of-art human body detection methods are studied. The focus is the challenge due to the human non-rigid body as well as complicated appearance and gesture.3. Based on the upper research and our specific problem. According to the various appreance, gesture and view angle of human body, the practicability of upper methods is analyed and their drawbacks are pointed out. And the probability optimization methods are introduced into our detection process, we interplayed between...