A Novel Algorithm for Diagnosis of Thin Basement Membrane Nephropathy

Abstract

In this paper we have made an algorithm to diagnose the thin basement membrane nephropathy. The idea of our algorithm is based on content based image retrieval and Hough transform. The diagnosis of this disease is depending on calculating the membrane thickness to know whether it is normal or abnormal. The traditional way for calculating the thickness is by manually enlarging the pictures for more than 5 thousand times before calculating the thickness, so we suggest an automatic algorithm to detect the membrane in the pictures then calculates the thickness. Firstly, a database of the membrane shapes will be build by dividing the original image of size 512512 pixel into sub images of size7070 pixel, the sub image that contain membrane will be considered, other parts will be ignored. Then, the sub image that contain the membrane will be enhanced and converted into binary image to detect the edges, Hough transform and line detect method are used to detect the surface of the membrane by drawing lines on the surface of the membrane, by applying orthogonal line on two lines that lies on the corresponding membrane surface, we then calculated the distance between two lines by using Euclidian distance. Compared with the manual procedures, our algorithm proves easy to use and can work round the clock.