A gray-level clustering reduction algorithm with the least PSNR

Gray-level clustering is an important procedure in image processing, which reduces the gray-level intensity of an image. In order to display a high gray-level image on low gray-level device screen, a good gray-level clustering reduction algorithm is necessary to complete this task. Based on the mean values and standard deviations of image histogram within different sub-intervals, a recursive algorithm for the gray-level reduction is proposed in this paper. It divides the image histogram into different sub-intervals recursively until the difference between original image and clustered image within given thresholds are reached. We experimented our proposed algorithm in comparison with other state-of-the-art algorithms on different high gray-level images. Our experimental results show our proposed algorithm outperformed others’ in terms of high visual quality of clustered images and computational inexpensiveness.