PSO based Gabor wavelet feature extraction method

In this paper, 2D continues Gabor wavelets are adopted to realize feature extraction. By optimize Gabor wavelet´s parameters of translation, orientation, and scale to make it approximates a local image contour region. The method of Sobel edge detection is used to get the initial position and orientation value of optimization in order to improve the convergence speed. In the wavelet characteristic space, we adopt PSO (particle swarm optimization) algorithm to identify points on the security border of the system. Comparing to the LM algorithm, it can ensure reliable convergence the target, which can improve convergent speed; the time of feature extraction is faster. By test in low contrast image, the feasibility and effectiveness of the algorithm are demonstrated by VC++ simulation platform in experiments.