An adaptive coding for singular value decomposition (SVD) samples is described in order to reduce the amount of transmitted data for monochrome images. This method is compared to the discrete cosine transform in terms of mean square error. Low bit rates, as 0.8 bits/pixel, are attempted with a good visual quality and low RMS error (0.087%). The visual and mathematical results of the transmission errors are shown and given.
Published in:
Communications, IEEE Transactions on
(Volume:27
,
Issue:
8
)
Date of Publication: Aug 1979