作者: Tom Mathes , Justus H. Piater
DOI: 10.1007/11861898_52
关键词:
摘要: We present an approach to non-rigid object tracking designed handle textured objects in crowded scenes captured by non-static cameras. For this purpose, groups of low-level features are combined into a model describing both the shape and appearance object. This results remarkable robustness severe partial occlusions, since overlapping unlikely be indistinguishable appearance, configuration velocity all at same time. The is learnt incrementally adapts varying illumination conditions target thus applicable any kind Results on real-world sequences demonstrate performance proposed tracker. algorithm implemented with aim achieving near real-time performance.