King Abdulaziz University Hospital Capsule dataset: A novel small-bowel endoscopic image repository from Saudi Arabia

Abstract

Wireless Capsule Endoscopy (WCE) has fundamentally transformed diagnostic methodologies for small-bowel (SB) abnormalities, providing a comprehensive and non-invasive gastrointestinal assessment in contrast to conventional endoscopic procedures. The King Abdulaziz University Hospital Capsule (KAUHC) dataset comprises annotated WCE images specifically curated for Saudi Arabian residents. Comprising 10.7 million frames derived from 157 studies, KAUHC has been classified into Normal, Arteriovenous Malformations, and Ulcer categories. Following the application of specific inclusion and exclusion criteria, 3301 labeled frames derived from WCE 86 studies were identified. Upon admission of patients, the data collection phase of KAUHC was initiated, involving the administration of the OMOM capsule and the use of the OMOM recording device for video documentation. A thorough evaluation of these recordings was undertaken by multiple gastroenterologists to identify any pathological abnormalities. The identified observations are subsequently extracted, categorized, and prepared for validation using Machine Learning (ML) classifiers. The dataset aims not only to address the scarcity of annotated endoscopic imaging resources in the Middle East but also to advance the development of diagnostic tools for ML applications in SB abnormalities and exploratory research on gastrointestinal diseases.

Keywords

• Classification
• Small-bowel mucosa
• Arteriovenous malformations
• Ulcer
• Machine learning