Turkish Journal of Electrical Engineering and Computer Sciences




Nephrons are the basic filtering units of the kidneys. Progression of chronic kidney disease (CKD) destroys nephrons permanently. Although there are many computing schemes suggested in recent years to identify CKD stages, no computing method has been suggested for identifying the nephron loss within kidney regions during CKD progression. In this paper, a novel pattern matching-based computation scheme is proposed to detect nephron loss in the kidney regions during CKD progression. We consider image registration (IR) with different transforms and a structural similarity index algorithm (SSIM) to match patterns of ultrasound images of kidney regions to identify the nephron loss. Simulation results show that the proposed scheme based on IR and SSIM algorithms detects almost 34% and 51% and almost 35% and 56% of the nephron damage in the cortex and medulla relative to a normal kidney, respectively. We extend the pattern matching scheme to identify CKD stages, as well. The proposed scheme based on IR and the SSIM algorithm can identify CKD stage with accuracy of 96% and 88%, respectively. The prediction accuracy of the proposed scheme for identifying CKD stage is comparable to that of the gray level co-occurrence matrix-based method, which is the best among the existing computing methods. However, the proposed scheme has advantages over the preexisting scheme, such as: the proposed method can be used to identify both CKD stage and the nephron loss in different kidney regions, and it identifies CKD stage without using a classifier. Therefore, the proposed pattern matching-based computing method is a better alternative to existing computing schemes for CKD stage identification.


Nephron, chronic kidney disease, ultrasound image, image registration, structural similarity index

First Page


Last Page