Multiple object tracking is being used for many applications nowdays such as automated surveillance, Robotics,self driving cars,medical and many more. There have been continuous improvements in existing state of art MOT(multiple object tracking) methods through many methods and global optimization techniques.This paper focuses on various MOT techniques and how to achieve speedup and efficiency using MOT methods.

As research in computer vision shifted from processing single, static images to the manipulation of video sequences the concept of movement recognition has become important. Movements are the most atomic primitives, requiring no contextual or sequence knowledge to be recognized. Movement is often addressed using either view-invariant or view specific geometric techniques. Here an attempt is made to develop a view-based approach to recognition of movement that is designed to support the direct recognition of the motion itself without reference to underlying static poses of the body. The basis of the representation is motion history image (MHI) -a static image where intensity is a function of the recency of motion in a sequence. Then recognition is performed by using both binary and scalar valued versions of the MHI as temporal templates to match against obtained instances of movement.