LEPset

LEPset is a large-scale EUS-based pancreas image dataset from the Department of Goenterology, Changhai Hospital, Second Military Medical University/Naval Medical University. This dataset consists of 420 patients and 3,500 images, and it has been divided into two categories (PC and NPC). We have invited experienced clinicians to annotate the category labels for all 3500 EUS images. Moreover, our LEPset also has 8,000 EUS images without any classification annotation.

1. After downloading the data set LEPset.zip, select the appropriate unzip file to extract it
2. After unzipping, there will be two folders: unlabeled and labeled
3. There are 8000 EUS images in the unlabeled folder and two folders in the labeled folder, NPC and PC, representing non-pancreatic cancer and pancreatic cancer respectively. 140 patients (1820 images) in NPC and 280 patients (1680 images) in PC
4. Unlabelled images can be used for pre-training of the model, and labelled images can be used for training and validation of the supervised model