Tracking multiple objects through occlusion with online sampling and position estimation

摘要: To track multiple objects through occlusion, either depth information of the scene or prior models such as spatial and smooth/predictable motion are usually assumed before tracking. When these assumptions unreasonable, tracker may fail. overcome this limitation, we propose a novel online sample based framework, inspired by fact that corresponding local parts in sequential frames always similar color texture features relative to centers objects. Experimental results illustrate proposed approach works robustly under difficult complex conditions.