Summary: To appear in Proceedings of the European Conference on Computer Vision (ECCV),
October 2008.
Tracking with Dynamic Hidden­State
Shape Models
Zheng Wu 1 , Margrit Betke 1 , Jingbin Wang 2 , Vassilis Athitsos 3 , and Stan
Sclaro# 1
1 Computer Science Department, Boston University, Boston, MA, USA
2 Google Inc.,Mountain View, CA, USA
3 Computer Science and Engineering Department, University of Texas at Arlington,
Arlington, Texas, USA
Abstract. Hidden State Shape Models (HSSMs) were previously pro­
posed to represent and detect objects in images that exhibit not just
deformation of their shape but also variation in their structure. In this
paper, we introduce Dynamic Hidden­State Shape Models (DHSSMs)
to track and recognize the non­rigid motion of such objects, for exam­
ple, human hands. Our recursive Bayesian filtering method, called DP­
Tracking, combines an exhaustive local search for a match between
image features and model states with a dynamic programming approach
to find a global registration between the model and the object in the
image. Our contribution is a technique to exploit the hierarchical struc­

Source: Athitsos, Vassilis - Department of Computer Science and Engineering, University of Texas at Arlington

Collections: Computer Technologies and Information Sciences