Abnormalities detection in kidney using multithreading technology

Kidney disease is one of the life threatening diseases prevailing among the humans. Most of the people die because of kidney diseases. It occurs due to the change which is occurring in the production of DNA cells (cancer), protein deficiency (nephritis) etc., In this paper, an automatic detection of the kidney diseases from CT abdominal images is proposed. First, the CT abdominal images are acquired and Region of Interest segmentation is performed for the kidney components, then the segmented image is preprocessed using color phase lab model which intends to remove the irrelevant noises and distinct the colors presented in image. The preprocessed image is further used for obtaining the infected region using Fuzzy C-Means clustering model. Feature selection of the segmented image is done by Gabor and PHOG features. With the use of Random Forest classifier, the segmented image is classified as abnormal and normal classes. A confusion matrix is estimated for analyzing the rate of prediction of the images to its relevant classes. Performance metrics such as True positive rate are estimated.