A Combinational Network for High Density Noise Removal and Texture Classification Using Back Propagation Network

Abstract

Image segmentation is an essential step in automatic pattern recognition and scene analysis problems. In this work, Artificial Neural Network (ANN) is used to segment the image. ANN based texture classification of an image of interest, is advanced technique providing rich information. This particular type of ANN is configured and trained in such a way that becomes capable of segmenting an image.The current work deals with the task where an object of interest is segmented out of a background for further processing. Along with this filtering is done by the modified decision based unsymmetrical trimmed median filter to avoid high density noise that cannot be removed by ordinary median filters. Hence the result of this overall network will be a high clarity texture classification.