Title :
A new class of median/mean based filters
Author :
Zeng, Bing ; Zhou, Hongbing ; Neuvo, Yrjo
Author_Institution :
Signal Process. Lab., Tampere Univ. of Technol., Finland
Abstract :
A new class of nonlinear filters which combines the thresholding operation with a median or mean filter is introduced. A simple but robust method to generate thresholds is proposed. Based on it, the new filters are analyzed for their white noise attenuation and abilities to eliminate impulse noise and preserve edges. Results are compared to those of the median, FIR median hybrid (FMH), and mean filters, showing several advantages of the new filters
Keywords :
digital filters; filtering and prediction theory; white noise; FIR median hybrid; edges; impulse noise; mean filter; median filters; nonlinear filters; robust method; thresholding operation; white noise attenuation; Adaptive filters; Attenuation; Filtering; Finite impulse response filter; Laboratories; Maximum likelihood detection; Noise robustness; Nonlinear filters; Signal processing; White noise;
Conference_Titel :
Circuits and Systems, 1990., Proceedings of the 33rd Midwest Symposium on
Conference_Location :
Calgary, Alta.
Print_ISBN :
0-7803-0081-5
DOI :
10.1109/MWSCAS.1990.140751
