Abstract Compression, transmission and archiving of digital images have become
necessities with recent developments in Internet communications and telemedicine.
An adaptive arithmetic coder has been developed first to archive the Visible
Human data set losslessly. While lossless compression is desired in many
applications, the attainable compression ratio is around only 2:1 and can
be improved to around 7:1 when segmented images are used. For higher compression
ratios wavelet-based lossy compression techniques that preserve significant
information are currently accepted even by some radiologists. Such compression
techniques can be applied to interactive use of the Internet to access
and retrieve the Visible Human (VH) images by progressive transmission
on a user-defined quality basis. Current MPEG-4 includes a wavelet based
compression technique that employs scalar quantization and is not optimal
but is suitable for fast progressive transmission. A recently developed
adaptive vector quantization technique optimized from the perspective of
rate distortion ( i.e. the distortion is relatively insignificant even
at very low bit rates) demonstrates potential strength in providing fast
yet high fidelity transmission of large color images such as the Visible
Human digital photograghic data.
