Gastroenterology and Artificial Intelligence: 2nd Annual Artificial Intelligence Summit | October 2020-Winning on Detection Classification and Navigation using AI

Winning on Detection Classification and Navigation using AI_Rivlin

Pdf Summary

This document discusses the importance of using artificial intelligence (AI) in diagnostics, specifically in the areas of detection, navigation, and classification. It highlights the high rate of medical errors, particularly in radiology and gastrointestinal imaging, and the need for improved accuracy in these areas. The document presents studies that demonstrate the potential of AI in detecting and classifying abnormalities, such as polyps, during procedures like colonoscopy. However, it also acknowledges that there are still challenges to overcome. One of these challenges is accurately determining the location and coverage during procedures, which requires advancements in techniques like landmark detection, ego motion analysis, and depth recovery. The document introduces solutions for solving coverage issues, including segment-based computation and real-time alerts. It also discusses unsupervised 3D reconstruction techniques for obtaining a 3D representation of the colon. Training methods for depth estimation, using both synthetic and real data, are presented, with promising results. The performance of AI systems in terms of accuracy is compared to that of doctors, showing that AI can outperform human experts. Finally, the document mentions the potential benefits of AI in patient history analysis, decision-making, skill assessment, and automatic reporting. Overall, while AI has shown promise in detection, classification, and navigation, further developments and improvements are needed to fully leverage its potential in improving diagnostic accuracy and patient outcomes.

Keywords

artificial intelligence

diagnostics

detection

navigation

classification

medical errors

radiology

gastrointestinal imaging

accuracy

polyps