Titlebook: Nonlinear Digital Filters; Principles and Appli I. Pitas,A. N. Venetsanopoulos Book 1990 Springer Science+Business Media New York 1990 Adap

阴郁

Algorithms and Architectures,rocessing applications. There are two solutions to the requirement for an increase of the speed of image processing routines. The first one is to construct fast algorithms for linear and nonlinear image processing. These algorithms can be relatively fast on general purpose computers for fast nonreal

Palpable

枯萎将要

愤世嫉俗者

Statistical Preliminaries,oped by statisticians in the last three decades and they have now reached maturity. In this chapter an introduction to robust statistics and order statistics will be given, as a mathematical preliminary to nonlinear filters. The interested reader can find more information in specialized books [1–5].

同来核对

Image Formation, are sensed through their radiant energy, e.g., light. Therefore, by its definition, image formation requires a .,an ., and a . Radiant sources can be of various kinds (e.g., white light sources, laser systems, X-ray tubes, thermal sources, even acoustic wave sources). Therefore, the physics of imag

纠缠,缠绕

Modicum

Digital Filters Based on Order Statistics,stics. This class includes a variety of nonlinear filters, e.g., .-filters, α-trimmed mean filters, max/median filters and median hybrid filters. Most of them are based on the well-known .-estimators. Therefore, they have excellent robustness properties. Others are modifications or extensions of the

PLIC

并置

Supplement

Polynomial Filters,mation problems, nonlinear filters arise in the case where the Gaussian assumption is not valid or the noise is not signal independent and/or additive. In the search for optimum signal processing systems, the task is to obtain general characterization procedures for nonlinear systems that retain at