A REVIEW ON MEAN SHIFT ALGORITHM BASED OBJECT TRACKING SYSTEM AND ITS HARDWARE IMPLEMENTATION

Abstract

Various attempts have been made to implement human vision in real-time computer vision. In most cases, however, this capability has been found to be too complex to implement on a practical systems. In short, human vision capability looks like a simple system but in reality it is difficult to model. One of the difficulties is the computational complexity required to process large amounts of data in real-time, and consequently, the design of real-time object tracking using vision remains a challenging problem. However, in this paper, we propose and implement hardware architecture for real-time object tracking system. We have chosen Mean-Shift algorithm for implementation. We will implement firstly this algorithm on MATLAB and simulated in different environments to verify the performance. After approving its robustness we implemented the algorithm on hardware.