The Evolutionary Approach Towards Swarm Optimization Algorithm Based Countenance Recognition

The countenance recognition presents a challenging problem in the field of image analysis and computer vision. Also, it received a good attention over the last few years. The applications of countenance recognition in the areas such as person tracking, surveillance, protection, entertainment, theft prevention, as easily as in the growth of human machine interfaces. It has been shown that precision in countenance is crucial to countenance recognition algorithms. An error of a few pixels in countenance will produce an important reduction in countenance recognition rates. In this paper, two approaches are proposed for use in countenance recognition of human. Firstly, Bacterial Foraging Optimization (BFO) and secondly, Particle Swarm Optimization (PSO). The Bacterial Foraging Optimization in which the features extracted from Linear Discriminant Analysis (LDA) and Independent Component Analysis (ICA) are optimized. The Particle Swarm Optimization (PSO), which optimizes the transform coefficients acquired from the Discrete Wavelet Transform (DWT) of the images. LDA and ICA are based on global impendence whereas DWT is performed on a block by block basis exploring the local impendence based features. The classifier performance and the length of the selected feature vector are pondering for performance evaluation using the Indian face database [35]. In this paper experimental results indicate that using Particle Swarm Optimization with DWT feature selection algorithm to proliferate transcendent countenance recognition outcome.