3D-EUS-PAF: A 3D Endoanal Ultrasound Image Dataset for Assessment of Perianal Fistulas and Abscesses

Published: 13 March 2024| Version 1 | DOI: 10.17632/vy5wv4t3zt.1
, Elahe Alimadadi,


3D-EUS-PAF dataset includes 160 Three-dimensional endoanal ultrasound images related to 34 patients. This dataset can be used by artificial intelligence (machine learning and deep learning algorithms) for diagnosis, classification, localization and segmentation of 7 classes that are available in the dataset: 1- Intersphincteric Fistula 2- Intersphincteric Abscess 3- Transsphincteric Fistula 4- Suprasphincteric Fistula 5- Perianal Fistula 6- Ischiorectal Abscess 7- Supralevator Abscess It also can be used for training medical students, interns, residents, specialists, and fellowships (Colorectal surgery). 3D-EUS-PAF is gathered and organized by a team including a professor of gastroenterology, an associate professor of colorectal surgery, a 3D-EUS specialist and a PhD, AI/ML Data scientist.



Shahid Beheshti University of Medical Sciences, Islamic Azad University Science and Research Branch


Gastroenterology, Medical Imaging, Colorectal Medicine, Machine Learning, Endoscope, Ultrasound, Image Classification, Colorectal Surgery, Medical Ultrasound, Fistula, Medical Image Processing, Deep Learning, Abscess