Automatic selection algorithm for region of interest of acne face image compression

Every day millions of skin and acne face images are generated all around the world and compression is the inevitability of all such types of images. Acne images suffer from the problem of poor compression ratios for obtaining high Peak Signal to Noise Ratio. Both, Peak Signal to Noise Ratio and Compression Ratio follow an inverse relationship. In this paper a new algorithm for acne image compression is proposed. The Proposed algorithm works in such a way that even a highly compressed image would retain all the mandatory information, since the region of interest (that is necessary for diagnosis) retains a significant quality and is compressed to an optimal level. Based on the image's Red and Green matrices, an automatic selection algorithm for the region of interest (patch) is introduced. The results of the proposed algorithm are compared with the conventional compression methods. A better peak signal to noise ratio of the instructive patch and overall compression ratios are achieved. Further the results are compared using different wavelet functions such as 'haar', 'db2'and 'bior1.3' wavelet. We also investigate proposed algorithms on other mother wavelets such as 'coiflet1' and 'symlet1'' and results are found to be superior.