KAUHC Dataset
Description
Wireless Capsule Endoscopy (WCE) has significantly altered the diagnostic approaches for small-bowel (SB) abnormalities, offering an exhaustive and non-invasive examination in contrast to traditional endoscopic techniques. This paper introduces the King Abdulaziz University Hospital Capsule (KAUHC) dataset, a novel WCE repository featuring annotated WCE images for Saudi Arabian residents. Encompassing 10.7 million frames sourced from 157 studies, KAUHC was categorized into Normal, Arteriovenous Malformations, and Ulcer classes. After careful inclusion and validation processes, 86 studies (i.e., 3301 labeled frames) were chosen. After the patient's admission, the data collection process of KAUHC began, wherein the OMOM capsule was administered, and the OMOM device was utilized for video recording. Subsequently, a thorough evaluation of these recordings by a gastroenterologist is undertaken to identify pathological abnormalities. The identified findings are then extracted, categorized, and prepared for validation. In terms of reuse potential, the dataset aims not only to mitigate the scarcity of labeled endoscopic imaging resources, particularly prevalent in the Middle Eastern region but also to foster the progression of diagnostic tools for Artificial Intelligence SB abnormalities and Gastrointestinal tract exploratory studies.
Files
Steps to reproduce
The collection process entails seven phases. Initially, the process commences with patient admission, preparation for GI tract screening, and the ingestion of the OMOM capsule. The second phase involves the capture, wireless transmission, and recording of a sequence of videos using the OMOM system. Then, Gastroenterologists conduct a manual review of the recorded videos in order to not only identify points of interest (POIs) but also choose the most relevant frames in terms of pathological abnormalities. Subsequently, the POI frames are exported, labeled, and submitted for the validation phase.