Automatic detection of pancreatic ductal adenocarcinoma in endoscopic ultrasound guided fine needle biopsy samples based on whole slide imaging using deep learning segmentation architectures.

Autor:	Podină, N., Udriștoiu, A. L., Ungureanu, B. S., Constantin, A. L., Georgescu, C. V., Bejinariu, N., Pirici, D., Burtea, D. E., Gruionu, L., Udriștoiu, Ș., Săftoiu, A.
Předmět:	DEEP learning PANCREATIC duct ENDOSCOPIC ultrasonography NEEDLE biopsy ADENOCARCINOMA IMAGE segmentation
Zdroj:	Endoscopy; 2024 Supplement 2, Vol. 56, pS95-S95, 1p
Abstrakt:	This article discusses the use of deep learning segmentation architectures to automatically detect pancreatic ductal adenocarcinoma (PDAC) in endoscopic ultrasound guided fine needle biopsy (EUS-FNB) samples. The study compares three U-Net architecture variants using two different datasets from medical centers in Craiova and Bucharest. The results show that the Inception U-Net model performed the best, with an accuracy of 97.82% for the Craiova Dataset and 95.70% for the Bucharest Dataset. The study concludes that the tested U-Net architectures provide excellent results for PDAC histological image segmentation. [Extracted from the article]
Databáze:	Complementary Index
Externí odkaz:	Zobrazit plný text záznamu