No label defined (Q4950143)

The authors provide a clearly written introduction in the basic notions and important applications of wavelets. The general ideas and approaches are presented in an intuitive way. E.g., in most cases the starting point for some topic is a question about how to treat a specific problem. The presentation makes the book in particular readable for engineering and applied mathematics students in their second year. It could serve as an appetizer for more advanced books or articles. The book splits in two parts. Part I discusses some theory and Part II is directed to applications. It starts with answering before hand a series of motivating questions: What are wavelets? For what they are made? Where are they useful? In particular two of the most important applications are shortly described: the compression of fingerprints and noise reduction algorithms. Chapter 2 reviews the standard material from discrete signal processing. As a rule the definitions are accompanied with hints to their relevance and importance, e.g., it is mentioned, that filters with linear phase are related to symmetric non-orthogonal wavelets. Continuous time signals are also treated. Chapter 3 proceeds with the discussion of filter banks representing one of the tools that lead to wavelets. In particular subsampling, alias cancellation, product filter, the design of filter banks, the Daubechies product filter and its factorization are described. Chapter 4 introduces the notion of multiresolution analysis, which expresses the idea of analysing a function at different resolutions or scales. Within this concept wavelets enter as the possibility to represent the difference between approximations at different scales. Chapter 5 presents several approaches to wavelets in higher dimensions. Besides separable tensor product wavelets quincunc and hexagonal wavelets are discussed. Lifting schemes which are related to charactizations of all compactly supported biorthogonal wavelet bases sharing the same low pass filter are the contents of Chapter 6. The continuous wavelet transform is studied in Chapter 7, describing, e.g., how it is able to reflect differentiability properties like local Lipschitz regularity. Part II starts in Chapter 8 with several examples for wavelet bases adding some general advice how to choose the suitable wavelet basis in certain applications. After introducing orthogonal and biorthogonal wavelets (Daubechies, Symmlets, Coiflets, spline wavelets) also the Meyer wavelets, Battle-Lemarié-wavelets and Chuis' semiorthogonal spline wavelets are presented. Chapter 9 shows constructions which allow an adaptation of the analysis to the signal (function) at hand: wavelet packets, local trigonometric bases, entropy and best basis selection. Compression and noise reduction including quantization and entropy coding are explained in Chapter 10. Further briefly sketched topics in the remaining Chapters 11-15 are fast numerical linear algebra algorithms (non-standard and standard form), differentiability and wavelet representation, feature extraction and implementation issues. A carefully written index finishes the book. Perhaps, more references could be added.