The DataScience@UB research group with the “Wireless Capsule Endoscopy Diagnosis by AI” led by Santi Seguí and Jordi Vitrià has been selected from the 5 finalists of the Mobile World Scholar Challenge 2019 in the topic Deep Learning.
The Mobile World Scholar Challenge is a competition, organized by the GSMA, aimed at university researchers who want to showcase their innovations in products and services that will revolutionize the technology industry.
The video describes the research made of the “Wireless Capsule Endoscopy Diagnosis by AI” project with the collaboration of Carolina Malagelada and Fernando Azpiroz, of the General Hospital of Vall d’Hebron and the Corporate Health company.
Capsule endoscopy is a procedure used to record images of the gastrointestinal tract for use in medical diagnosis. This procedure involves the ingestion of a small capsule, similar in shape to a standard pharmaceutical pill, that contains a tiny camera and an array of LEDs powered by a battery. The capsule takes a number of images per second, which are wirelessly transmitted to an array of receivers connected to a portable recording device carried by the patient. The use of video capsule endoscopy in the colon has been proposed as an alternative colorectal cancer-screening test.
The main objective of this applied research project has been to develop a scalable AI solution to significantly speed up the process of inspection of the gastrointestinal tract. By using Deep Learning techniques, a computer can inspect hundreds of thousands of images and select the ones with anomalies, just in a few minutes instead of a few hours.
The technological breakthrough of this solution is the reduction in time and costs of analysis for the diagnosis, as no professionals, proper facilities and other medical devices are needed.
Colorectal cancer screening is nowadays costly, invasive and labour-intensive, and deemed an unsuitable population-wide index screening tool. This Artificial Intelligence system democratises access to screening so that as many patients as possible can benefit from early detection of serious diseases.