Particle swarm optimized feature selection for retrieving compressed medical images

This Compressed medical image plays an important role as health centers move towards a complete computerization. Additionally the relative review is acquainted with decide the best method for medical image compression. Edge location strategies change unique pictures into edge pictures compensation from the progressions of gray levels in the picture. The work is especially focused on towards wavelet image compression utilizing Haar Transformation with a thought to minimize the computational necessities by applying diverse compression limits for the wavelet coefficients and these outcomes are acquired in division of seconds and subsequently to enhance the nature of the constructed image. Mutual information is an accepted comparison metric for medical image registration. Sobel edge detector and Gabor filter are used for feature extraction. Recurrent Neural Network is used for classification of images. Since feature selection is NP hard, a Particle Swarm Optimization (PSO)is proposed to optimize the feature subset selection. The results obtained are satisfactory.