Abstract | ||
---|---|---|
Interstitial fibrosis in renal biopsies has shown a good correlation to the presence of chronic kidney disease, and it is therefore quantified by pathologists in the diagnosis of the disease. In the previous work, the developed automatic quantification system for the interstitial fibrosis was presented. It was based on the segmentation of tubular structures. This paper advances the development of the system by expanding the set of identifiable structures to include glomerulus, arteries, and urinary casts. In particular, it investigates two methods of the glomerular detection, namely the Bowman's space search and rLADTree classification. The quantification results of the final system incorporating segmentation of all commonly seen structures have shown a quantification matching that produced by the ground truth to within 8.8%, while the glomerulus structure detection by rLADTree classification method outperformed its alternative due to its higher robustness in tolerating the variance in the glomerulus appearance. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1109/I2MTC.2017.7969716 | 2017 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) |
Keywords | Field | DocType |
renal fibrosis,medical image analysis,glomerulus segmentation,rLADTree | Fibrosis,Medical imaging,Image segmentation,Control engineering,Artificial intelligence,Pathology,Kidney,Urinary system,Computer vision,Segmentation,Kidney disease,Renal interstitial fibrosis,Mathematics | Conference |
ISBN | Citations | PageRank |
978-1-5090-3597-7 | 0 | 0.34 |
References | Authors | |
4 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Wei Keat Tey | 1 | 0 | 0.34 |
Ye Chow Kuang | 2 | 72 | 19.81 |
Joon Joon Khoo | 3 | 0 | 1.01 |
Melanie Po-Leen Ooi | 4 | 70 | 18.35 |
Serge N. Demidenko | 5 | 84 | 19.38 |