作者: Lifeng Liu , Stan Sclaroff
DOI: 10.1016/S0167-8655(02)00108-3
关键词:
摘要: Abstract An improved method for deformable shape-based object detection and segmentation is described. A pre-computed index tree used to improve the speed of template fitting. Simple shape features are as keys in a pre-generated model instances. coarse fine indexing scheme at different levels further speed. The approach demonstrated an improvement previously-reported template-based region system. Experimental results show that when trees used, speedup significant while accuracy maintained.