Results of our method on frames of a video sequence. Boundary delineation is shown in red, with green circles indicating recovered joint positions, and yellow points for shape context based initialization positions. The images demonstrate the method's ability to handle cases with an exceptionally unclear boundary between the arm and the torso.
See example movie.
More details and movies
Chris McIntosh, Ghassan Hamarneh and Greg Mori. Human Limb Delineation and Joint Position Recovery Using Localized Boundary Models. IEEE Workshop on Motion and Video Computing, 2007. [pdf]