Time-Frequency/Time-Scale Analysis

This highly acclaimed work has so far been available only in French. It is a detailed survey of a variety of techniques for time-frequency/time-scale analysis (the essence of "Wavelet Analysis"). This book has broad and comprehensive coverage of a topic of keen interest to a variety of engineers, especially those concerned with signal and image processing. Flandrin provides a discussion of numerous issues and problems that arise from a mixed description in time and frequency, as well as problems in interpretation inherent in signal theory.

Key Features
• Detailed coverage of both linear and quadratic solutions
• Various techniques for both random and deterministic signals

Audience: Students and practitioners of mathematics, computer science, and electrical engineering, especially image processing engineers; universities offering courses in image compression.

Originally published in French as Temps-frequence, this book is a multi-disciplinary introduction to the toos of time-frequency signal analysis (often referred to as wavelet analysis). It serves as an invaluable reference for anyone concerned with the application of these techniques to signal processing.

Although Fourier analysis provides a natural environment for the description of stationary signals, most real signals are nonstationary. Indeed, most often an integral part of the information contained in a signal can be found in the nonstationary property itself (e.g., start and termination of events, drifts, ruptures, modulations). This book provides a coherent and self-contained presentation of a broad panorama of methods for analyzing nonstationary signals, an area which has recently undergone intensive development in signal processing. Special emphasis is given to non-parametric approaches, with particular focus on quadratic energy distributions. This book offers a discussion of problems that emerge from a mixed description in time and frequency, and it comprises a detailed construction of the principle classes of admissible solutions (short-time Fourier, wavelets, Wigner-Ville, Cohen's class, affine distributions, etc.). It also addresses numerous issues of interpretation related to such representations and shows how to reformulate several important problems in signa theory by utilizing joint time-frequency distributions.

Booknews

Originally published in France as quence/>, this text is a multi-disciplinary introduction to applying the techniques of time- frequency signal analysis (or wavelet analysis) to signal processing. Special emphasis is given to the analysis of information embedded in nonstationary processes (e.g. drifts, ruptures, and modulations). Non-parametric approaches are given special treatment, particularly with regard to quadratic energy distributions. Annotation c. by Book News, Inc., Portland, Or.

Patrick Flandrin is a Research Director of the CNRS (Centre National de la Reserche Scientifique). He holds a position at the Ecole Normale Superior in Lyon, where he leads the group "Traitement du Signal" (Signal Processing), which is part of the Physics Department. He is also Associate Director of the federative CNRS structure, information, Signal, Images et Vision (ISIS), and has served on the editorial boards of several journals. He was awarded the Philip Morris Scientific Prize in 1991.