Image data is compressed by decomposing the image data into a plurality of
resolutions or sized by wavelet decomposition, followed by compression of the resulting
data sets. Adaptive Huffman code compression is optimized for compressing the data
sets, predictive errors for a low frequency band and the wavelet transformed values
for high frequency bands. The image data is thus compressed losslessly and in multiple
resolutions such that desired resolution images can be reconstructed and transmitted
in accordance with available-bandwidth and viewing capabilities.