3D Medical Image Referring Segmentation Dataset (M3D-RefSeg),
consisting of 210 3D images, 2,778 masks, and text annotations.
3D medical segmentation is one of the main challenges in medical image analysis. In practical applications,
a more meaningful task is referring segmentation,
where the model can segment the corresponding region based on given text descriptions.
However, referring segmentation requires image-mask-text… See the full description on the dataset page:
https://huggingface.co/datasets/GoodBaiBai88/M3D-RefSeg.
