Medical imaging has started to take advantage of digital technology, opening the way for advanced medical imaging and
teleradiology. Medical images, however, require large amounts of memory. At over 1 million bytes per image, a typical
hospital needs a staggering amount of memory storage (over one trillion bytes per year), and transmitting an image over a
network (even the promised superhighway) could take minutes--too slow for interactive teleradiology. This calls for
image compression to reduce significantly the amount of data needed to represent an image. Several compression
techniques with different compression ratio have been developed. However, the lossless techniques, which allow for
perfect reconstruction of the original images, yield modest compression ratio, while the techniques that yield higher
compression ratio are lossy, that is, the original image is reconstructed only approximately. Medical imaging poses the
great challenge of having compression algorithms that are lossless (for diagnostic and legal reasons) and yet have high
compression ratio for reduced storage and transmission time. To meet this challenge, we are developing and studying
some compression schemes, which are either strictly lossless or diagnostically lossless, taking advantage of the
peculiarities of medical images and of the medical practice.
In order to increase the Signal to Noise Ratio (SNR) by exploitation of correlations within the source signal, a method of
combining differential pulse code modulation (DPCM) is presented.