Medical SAM 2 Sample Medical Segmentation Dataset

This dataset was created by researchers at the University of Oxford in 2024 in the paper “Medical SAM 2: Segment medical images as video via Segment Anything Model 2", the sample data used includes REFUGE ophthalmology medical image dataset(for 2D cases) and BTCV abdominal multi-organ image segmentation dataset(for 3D cases), both of which were pre-trained by the research team.

The REFUGE dataset contains 1,200 color fundus photographs analyzed for glaucoma, divided into ground-truth segmentations and clinical glaucoma labels.

The BTCV dataset comes from the organ segmentation challenge task in the Medical Universal Segmentation Competition. The input image contains 1 channel, namely the CT electron density image, and the label image contains 14 channels (segmentation of 13 organs plus background points).