Gaussian mixture grayscale image enhancement algorithm based onparticle swarm optimization

A Gaussian mixture model(GMM) based grayscale image enhancement algorithm using particle swarm optimization(PSO) is proposed. The algorithm uses GMM to build a model for gray level histogram of the input image, and uses the significant interaction points of Gaussian components in the model to partition the histogram into a certain amount of sub intervals. Then, according to mapping function, the gray value in each interval is transformed to appropriate output interval. The enhanced image is generated by transforming output interval data according to PSO optimized parameters. Experimental results show that image visual effects generated by Gaussian mixture gray image enhancement algorithm with PSO are better. After image processing of original image and texture rich image, the information entropy of enhanced image is 4.746 6 and 7.952 6 respectively, the average gray gradient is 6.970 6 and 37.386 1.