Glomerulus diameter and Bowman's space width in renal microscopic images indicate various diseases. Therefore, the detection of the renal corpuscle and related objects is a key step in histopathological evaluation of renal microscopic images. However, the task of automatic glomeruli detection is challenging due to their wide intensity variation, besides the inconsistency in terms of shape and size of the glomeruli in the renal corpuscle. Here, a novel solution is proposed which includes the Particles Analyzer technique based on median filter for morphological image processing to detect the renal corpuscle objects. Afterwards, the glomerulus diameter and Bowman's space width are measured. The solution was tested with a dataset of 21 rats' renal corpuscle images acquired using light microscope. The experimental results proved that the proposed solution can detect the renal corpuscle and its objects efficiently. As well as, the proposed solution has the ability to manage any input images assuring its robustness to the deformations of the glomeruli even with the glomerular hypertrophy cases. Also, the results reported significant difference between the control and affected (due to ingested additional daily dose (14.6mg) of fructose) groups in terms of glomerulus diameter (97.40±19.02μm and 177.03±54.48μm, respectively).
Keywords: Glomerular hypertrophy; Image analysis; Medical imaging; Particles analysis algorithm; Renal diseases.
Copyright © 2015 Elsevier Ireland Ltd. All rights reserved.