RGB Channel Analysis for Glomerulus and Proximal Tubule Detection in Kidney Histology Image
Ima Kurniastuti, Ula Nur Rahmatin, Firman Yudianto, Tri Deviasari Wulan · 2021
Segmentation in digital pathology, especially kidney histology images, is still rarely done. The detection of parts of the kidney is done manually by an experienced laboratory which takes a long time. This paper proposed method of extracting color features in glomerulus and proximal tubule as the initial stage of segmentation. Data were kidney histology images. RGB color feature extraction is done by utilizing histogram. Histogram give information about highest frequency of intensity. Data analysis was statistically using grouped frequency distributions to obtain range number of RGB. The result was RGB in kidney histology image can be extracted and comparison of color components in RGB show that all component overlap in the glomerulus while in proximal tubule, blue component can be used as feature differentiator. The conclusion was RGB channel cannot be used as a distinguishing feature in the segmentation of parts of the kidney, especially the glomerulus and proximal tubule.