The authors describe efficient, fast numerical algorithms for computing wavelet transform representations. Methods presented include modified lattice and transform domain computation. The implementation of the wavelet transform is computationally intensive, since the number of computations required increases with the number of octaves. Various methods are applied to reduce the number of operations necessary for calculations of wavelet representation. Computational complexity per octave is compared to several algorithms described here. Modified lattice structure results in better efficiency for shorter wavelets, while transform domain filtering with incorporated sampling rate conversion leads to computationally efficient implementation of wavelet coefficients for longer wavelets
Published in:
Circuits and Systems, 1992. ISCAS '92. Proceedings., 1992 IEEE International Symposium on
(Volume:2
)
Date of Conference: 10-13 May 1992