Wavelet Based Compression of Acne Face Images with Automatic Selection and Lossless Compression of Acne Affected Region

Abstract

Medical image compression is inevitable part of medical research centers and hospitals. In this paper, compression of acne face images is considered by exploiting the fact that such images normally have some region of interest (ROI) space that contains acne and other space (without acne) as non region of interest. For proper diagnosis and treatment of acne, compression should be performed in such a way that no information loss results for the acne affected region. This paper proposes a new algorithm, in which the acne affected region is automatically selected using K-means clustering and then compressed minimally whereas relatively higher compression is applied on the non region of interest using wavelet transform in RGB colour space. Using this algorithm good Compression Ratio (CR) upto 8-14 is achieved without degradation in image quality of the acne affected region.