The aim of digital image coding for compression is to minimize memory for storage and/or bandwidth for transmission. In the case of grayscale images, compression is typically accompanied by some distortion or loss of information in the reconstructed image. Generally, the larger the distortion, the better the compression; a balance between these competing trends has to be achieved in any practical coding design. Different coding schemes can result in distortions which are perceived differently by viewers. These various schemes can also give different compression rates, but rate comparisons among them are hampered by the lack of a good measure of image fidelity consistent with subjective appreciation of image quality by the human viewer. A problem that is simpler, but still useful, is the study of the rate-distortion tradeoff for a particular type of distortion such as that produced by a particular class of coders.
