Performance Study on Facial Image Compression Using Master Code

In this paper, face images are studied in the context of detection and recognition. In the literature, the problem of compressing human frontal facial images has been addressed. The images that are dealt are passport-type photos namely, full face, frontal view, plain background, no dark glasses, without hates and other non-standard clothing. Compression of Facial Images by using IDL is carried out in this paper. The facial images dealt with are passport-type photos-full face, frontal view. Three different photos of the same person are stored. A single code book is generated and compression is carried out. Appreciable PSNR values are obtained for different block sizes. This paper introduces a method to compress and decompress the facial image, so that it is visually appealing, at a quality sufficient to an uncertainty in visually identifying a given subject.