Patch based frame work for occlusion detection in multi human tracking

Nowadays video surveillance system becomes popular to ensure security at major areas. Current video surveillance system analyzes the tracking of multiple human in a video. The problem of detecting moving targets in crowded scenes is one of the most challenging topics in such systems. Occlusion is said to be one object hides another during multiple human tracking in a series of frames or in one frame. Occlusion detection and handling is an important phase for such systems. The main objective of the proposed work is to detect occlusion. In the proposed work, patches in each object are used to identify the parts which are going to be occluded. This system can identify the occlusion between humans in a scene and can identify the occluded parts of each object. Our ultimate goal is to develop a surveillance system that is free from occlusion.