Texture descriptions and classification for pathological analysis of cancerous colonic mucosa

Image analysis for the classification of normal and cancerous colonic mucosa is reported. Pathology samples were taken from human colon mucosa, and 44 normal and 58 cancer images captured to computer via an optical microscope with a CCD camera. Texture analysis was performed using fractal dimension, entropy and correlation. Using non-parametric classification, fractal dimension improved the classification accuracy from 88% to 94% in comparison with the combined entropy and correlation analysis