Open AccessArtificial intelligence applications in gastroenterology: steps ahead
Author(s) -
Stefan L. Popa,
Abdulrahman Ismaiel
Publication year - 2022
Publication title -
medicine and pharmacy reports
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.35
H-Index - 16
eISSN - 2668-0572
pISSN - 2602-0807
DOI - 10.15386/mpr-2513
Subject(s) - convolutional neural network , computer science , artificial intelligence , artificial neural network , medical imaging , capsule endoscopy , endoscopy , magnetic resonance imaging , medical physics , image processing , deep learning , radiology , medicine , image (mathematics)
Artificial intelligence (AI) applications are used in gastroenterology for automatic imaging diagnostic methods such as ultrasonography, computer tomography, magnetic resonance imaging, but also in endoscopy, capsule endoscopy and biopsy followed by automatic digital pathology evaluation. The accuracy of AI-based systems is superior to human expertise. Furthermore, in reality, a very small percentage of the patients are being investigated by a human expert in endoscopy, so implementing AI in this investigation would only increase the diagnostic accuracy. The existence of an unimaginable number of digital images and different types of medical information made possible the analysis and training of convolutional neural network (CNN), which consists of multilayers of artificial neural networks (ANN) with step-by-step minimal processing, creating a fundamental resource for any AI-based system to learn by itself how to automatically perform medical tasks, which were performed only by human experts in the past. The main objectives for AI applications used in gastroenterology are to improve the medical procedures with enhanced precision, to reduce the number of medical errors and to perform repetitive tasks.