Particle Swarm Optimization in Foreground Segmentation for Live Video

The goal of foreground segmentation is to extract the desired foreground object from input videos and matting problem means the problem of extracting the foreground object accurately. Over the years, there have been significant amount of efforts on how to extract objects from live videos. The support vector machine (SVM) classification is an active research area which solves classification by finding the best hyperplane that separates all data points of one class from those of the other. In many problems of classification, the performances of SVM are often evaluated by the rate of error depends on the optimization method adopted to label the unknown pixels along the boundary. Here, among the methods of different optimization methods, selected the method of PSO (Particle Swarm Optimization) which makes it possible to optimize the performance of classifier and it will enhance the accuracy of the foreground object segmentation of live videos. PSO is a population based stochastic search process, shaped after the social behavior of a bird flock. It is similar to flock of birds migrating towards some destination, where the intelligence and quality lies in the co-operation of an entire flock. So over the iteration, a group of variables have their values adjusted closer to member whose value is closest to the target at any given moment. Foreground segmentation is addressed in the application of new background substitution and shown to create convincingly high quality composite video.