Research And Development Of Human Motion Analysis

Human motion analysis based on monocular video sequences is an important research area which combines computer vision and graphics techniques. The focus of the research is in detecting and tracking figures in monocular videos, capturing and reconstructing the human motion, and then describing and interpreting the human behavior. This research has various applications in human animation, computer game, virtual reality and augmented reality, human-computer interaction video surveillance, sports video analysis, computer-aided clinical diagnosis and so on.This thesis focuses on the problems of tracking human and capturing three dimensional (3D) human motion in monocular videos. This work investigates on tracking the articulated motion and evaluating the relative three dimensional motion of the figure in the form of articulated skeleton chains in monocular videos. Approaches of tracking and capturing human motion in monocular videos are proposed. This research is anticipated to help obtain the abundant motion data from plenty of videos including movie, TV and sports videos, and enable the myriad activities of human created in virtual world.The contributions of the thesis are listed as follows:1) An articulated human skeleton model suitable for human motion tracking and 3D motion reconstruction is proposed. Based on the basic human skeleton, the thesis proposed an articulated skeleton human model suitable for human motion tracking and 3D motion reconstruction. Compared with the model proposed by previous researcher, this model simplified the motion tracking method and 3D motion reconstruction, reducing the dependence of existing knowledge. Since 3D articulated skeleton human model is as same as two dimensional (2D) one, it is convenient to estimate 3D human motion reconstruction after obtaining the 2D joint coordinates.2) A novel approach on human motion tracking in monocular videos is proposed. By using the geometrical human skeleton model, joint coordinates can absorb on the skeleton through tracking algorithm. If error occurs in the tracking process, it can be fixed automatically. This novel approach uses the motion data obtained from the training video sequences, and no 3D motion capture data is required, getting rid of the dependence of expensive motion capturing equipments and no special background requirement is needed.3) A method on 3D reconstruction of the human motion form monocular videos is proposed. As camera parameter can not be obtained from monocular video sequences, a scaled orthographic projection model is employed to evaluate the relative 3D motion of the figure. The relative lengths oft he segments are obtained form the appropriate frames, and the articulated human skeleton model is customized with them. The 2D joints in first frame are obtained manually. Then, the process of the new approach is implemented through parameter evaluation and optimization in each frame. The method doesn't require special equipments except for some user interaction. It is a robust method to estimate 3D human motion, meeting the needs of human animation.The proposed approaches provide convenient and practical ways of human motion tracking and motion reconstruction in monocular videos. The objective of obtaining human motion data from the plenty videos such as movies and sports videos is realized at the beginning steps. The contributions made in this thesis provide a step of progress on the way towards robust and accurate tracking and capture of human motions in monocular video sequences. Suggestions are proposed in the last part of the thesis.