5. Acknowledgments

We wish to thank Scott Nations for his many helpful comments on an early version of this paper. A theory for multiresolution signal decomposition: the wavelet representation, " IEEE Trans. Approximated Ocean Area Land (a) (b) Figure 7: Land and ocean areas in each data frame. (a) Ocean area in original data (white area). (b) Approximation of the ocean area using 88 blocks. This approximated version entirely covers the original ocean area. h *2 + ƪǒ1)cosa)sina)· (1 *cosb*sinb))2cosbsina]ń4 h *1 + ƪǒ1*cosa)sina)· (1)cosb*sinb)*2cosbsina]ń4 h 0 + [1)cos(a*b))sin(a*b)]ń2 h 1 + [1)cos(a*b)*sin(a*b)]ń2 where *p v a, b v p (9) parameter space for the minimum MSE and compared it with the MSE generated by the spline biorthogonal wavelet compression algorithm. Table II gives the results of some typical signals in our application. From these data, we concluded that although the lengths of the spline biorthogonal WT filters are less than 6, they perform the same or better than the optimized 6–tap WTs in a scheme where both forward and inverse transforms use the same filters.