Comparison of SPIHT, Classical and Adaptive Lifting Scheme for Compression of Satellite Imageries

The signals that are encountered in practice are not smooth signals and classical wavelet transforms cannot deal with the discontinuities encountered in the signals. Such singularities tend to give rise to large coefficients in their proximity, which is undesirable for signal compression. To overcome such problems one can consider the local variance during the decomposition of the signal. There are various ways to build adaptivity into the decomposition of a signal. The best algorithm, selects a wavelet basis by minimizing a concave cost function such as the entropy. In such an approach, the filter coefficients are fixed for entire block of data as the optimization criterion is global. Here, the decompositions are considered where the filter coefficients vary locally, taking into account of local signal variations. The approach taken by Chan and Zhou [1] suggests that instead of changing the filter coefficients, the input signal is changed in the proximity of discontinuities through an extrapolation procedure. By registering these changes, the original signal can be recovered at synthesis level. By extending the approach of Chan and Zhou in the present work, the SPIHT, Classical Lifting scheme and Adaptive Lifting schemes are analyzed for achieving better compression ratio and PSNR for satellite Rural and Urban imageries. The results are presented in the paper.