Two‐dimensional neural hybrid filters using adaptive windows and layered median filters
Mitsuji Muneyasu, Takahiro Maeda, Takao Hinamoto · Electronics and Communications in Japan (Part III Fundamental Electronic Science) · 1996
Abstract A new 2‐D neural hybrid filter tha combines a layered median filter with adaptive‐window linear filters is described. the filter can eliminate both impulsive and Gaussian noise and has good edge preservation characteristics, owing in particular to the 3‐point layered median filter design. the size of the adaptive windows is maximized in regions where the neighboring pixel values differ relatively little from that at the working point. the design yields effective noise removal without edge degradation, and learning by the neural network allows the filter to adapt to image characteristics.