Authors: Carlo Biffi, Giulio Antonelli, Sebastian Bernhofer, Cesare Hassan, Daizen Hirata, Mineo Iwatate, Andreas Maieron, Pietro Salvagnini, Andrea Cherubini

Published on: March 04, 2024

Impact Score: 8.4

Arxiv code: Arxiv:2403.02163

Summary

• What is new: Introduction of the REAL-Colon dataset, a large-scale, high-quality collection of real-world, full-resolution colonoscopy video frames, annotated by experts.
• Why this is important: Current public datasets for colonoscopy consist mainly of low-resolution images or clips, inadequately representing the complexities of real procedures.
• What the research proposes: The REAL-Colon dataset offers 2.7M video frames from real procedures with comprehensive annotations and clinical data, enhancing AI research in colonoscopy.
• Results: Provides a unique, high-quality resource for developing and testing more accurate AI models for detecting and diagnosing colon polyps.

Technical Details

Technological frameworks used: nan

Models used: nan

Data used: REAL-Colon dataset with 2.7M native video frames, 350k annotations, and comprehensive clinical data.

Potential Impact

Healthcare providers, endoscopy equipment manufacturers, and AI research companies focusing on medical diagnostics and treatment.

Want to implement this idea in a business?

We have generated a startup concept here: ColoAI.