High Definition Video Segmentation Techniques-A Review
R. S. Sabeenian, S. Rama Lavanya · International Journal of Computer and Electrical Engineering · 2013
Real time segmentation of moving regions in image sequences is a fundamental step in many vision systems including automated visual surveillance human-machine interface and very low bandwidth telecommunications. Background identification is a common feature in many video processing systems. One of the most important back ground identification algorithm is the Gaussian Mixture Model algorithm (GMM).On implementation of the Gaussian mixture model on FPGA results in Reduction of the processing capability of the overall system. Trainable Segmentation is adapted to improve the processing capability. After analyzing and evaluating the performance we conclude with several promising directions for future research.