Part of the Endoscopic Vision Challenge



For the diagnosis of gastroscopic images, the conventional way is to employ visual interpretation of an expert physician. With the development of computer technology, a warning/support system to classify image regions as normal or abnormal will assist medical experts by reducing their manual workload and improving the accuracy of medical diagnosis.


In our sub-challenge, we will focus on comparing different abnormal detection methods for recognizing the abnormal regions from gastroscopic images. We will address the abnormal detection for gastroscopic images with different abnormal patterns, such as gastritis, cancer, ulcer and bleeding.


Overall, participants are invited to submit results for the following task:


Detection of abnormalities: Can you detect the abnormal regions in a gastroscopic image?