Liver Image Analysis using Color and Texture Descriptors

Microscopic imaging is increasingly becoming useful in analyzing problems and processing techniques in digital image processing. Liver failure is caused by improper functioning of hepatocytes that leads towards various liver diseases if not detected earlier. The automated system is proposed for normal and suspected liver samples detection. Features selection is based on texture and color properties of microscopic images. In classification, SVM classifier with Radial basis function is used to correctly classify the abnormal liver images contained hepatocyte cells. Performance of proposed system is tested on mice liver dataset and 83% accuracy is achieved.