Abstract
Publisher Summary A system that reconstructs 3D human movement from a stream of 2D input images is outlined in this chapter. The system finds the 3D shape and motion of the human body by fitting a simple skeleton model to the markers found in the 2D image. The fitting is done using a nonlinear optimization approach. This optimization is essentially difficult because in-depth information cannot be obtained from 2D images. In order to overcome this problem, an inverse analysis technique that takes account of the following four kinds of a priori information is applied. First, the human body can be exhibited by versatile skeleton model. Second, joint angles have their limits. Third, human motion is smooth. Fourth, human keeps one's balance in slow motion. This chapter presents a reconstructed human movement from real video images by this method in order to demonstrate its applicability.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
