A bounded distortion compression scheme for hyper-spectral image data

Presents a new approach for nearly lossless compression of multispectral image data that exploits both spectral and spatial correlations in a simple and adaptive manner. What the authors have described is just one choice of predictor, re-ordering and encoding. A number of alternatives can be used. Implementation results with a few different choices schemes are currently under investigation and will will be described at a later date. Also, the authors need to make more detailed comparisons of compression performances obtained with other lossy and nearly-lossless schemes given in the literature