Endoscopy Recommender System

Endoscopy Recommender System is one of our first machine learning research projects. It focuses on gastrointestinal anomaly recognition on endoscopy videos. Project includes:

• database of few hundred endoscopic videos and few thousand images, annotated by specialists
• advanced image preprocessing algorithms
• many image classifications algorithms, allowing for informative frame recognition, disease classification (cancer, ulcer, polyp, etc.), bleeding detection and more
• machine learning and image processing research platform

During implementation of the project, many scientific publications were created, see here.

Algorithms created in ERS are used in MedEye application on KASKADA platform.