A finite wavelet decomposition method
Sumio Watanabe · Electronics and Communications in Japan (Part III Fundamental Electronic Science) · 1997
A method is proposed for approximate expansion of a given function in terms of a finite sum of wavelets based on the solvable model in nonlinear approximation problems. Its applications to multiresolution analysis and data compression are discussed. Unlike the conventional function decomposition by wavelets, the proposed algorithm is distinguished by its ability to automatically extract only the wavelets needed for description of the object function and is hence considered useful for function approximation and data compression. The method also has an advantage useful for applications in that the generation of functions with different resolutions can be carried out easily. The practicability of the method is confirmed by an application test to several functions. © 1997 Scripta Technica, Inc. Electron Comm Jpn Pt 3, 80(7): 1–10, 1997