ERS - sporedata/researchdesigneR GitHub Wiki

General description

The ERS (Endoscopic Reference System) dataset is a new multi-label comprehensive image dataset from flexible capsule endoscopy, colonoscopy, and endoscopy. ERS contains around approximately 115,000 and precisely 6000 labeled frames from endoscopy videos, 22,600 approximate and 3600 precise segmentation masks, and 1.23 million unlabeled frames from capsule and flexible endoscopy videos, all derived from 1520 videos of 1135 patients.

The ERS dataset is developed to provide a diverse set of images captured during various endoscopic procedures, which include examinations of the gastrointestinal (GI) tract, such as gastroscopy and colonoscopy. Its primary goal is to support ML algorithms in detecting abnormalities like polyps, tumors, ulcers, or inflammation, potentially enhancing diagnostic accuracy and efficiency.

Related publications

• ERS: a novel comprehensive endoscopy image dataset for machine learning, compliant with the MST 3.0 specification

Data access

For more information on the ERS dataset, visit https://arxiv.org/abs/2201.08746