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

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.

3D-EUS-PAF 数据集包含34例患者的160幅三维肛内超声（Three-dimensional endoanal ultrasound）图像。本数据集可用于依托人工智能（机器学习与深度学习算法）开展数据集中7类病变的诊断、分类、定位与分割任务，具体病变类别如下：
1. 括约肌间瘘（Intersphincteric Fistula）
2. 括约肌间脓肿（Intersphincteric Abscess）
3. 经括约肌瘘（Transsphincteric Fistula）
4. 括约肌上瘘（Suprasphincteric Fistula）
5. 肛周瘘（Perianal Fistula）
6. 坐骨直肠窝脓肿（Ischiorectal Abscess）
7. 肛提肌上脓肿（Supralevator Abscess）

此外，该数据集还可用于培训医学生、实习医师、住院医师、专科医师以及结直肠外科进修医师。
3D-EUS-PAF 数据集由一支包含消化内科教授、结直肠外科副教授、三维肛内超声专科医师与一名人工智能/机器学习数据科学博士的团队采集并整理。