Title :
Optimal edge-preserving hybrid filters
Author_Institution :
Dept. of Electron., Shandong Univ., Jinan, China
fDate :
11/1/1994 12:00:00 AM
Abstract :
Based on the least mean square error criterion at the edge position, a new algorithm is presented to obtain a class of optimal edge-preserving hybrid filters that include the Lee (1980) additive filter and the gradient inverse weighted filter. As a result, we show that the gain of the Lee additive filter is not optimal and derive the optimal form of this filter. In addition, some examples are given
Keywords :
circuit optimisation; digital filters; edge detection; filtering theory; image processing; least mean squares methods; nonlinear filters; algorithm; gradient inverse weighted filter; least mean square error criterion; optimal Lee additive filter; optimal edge-preserving hybrid filters; Additives; Gaussian noise; Image processing; Mean square error methods; Nonlinear filters; Signal processing algorithms; Smoothing methods; White noise;
Journal_Title :
Image Processing, IEEE Transactions on
