Kimokcheon/

FAME_benchmark

FAME is a benchmark for zero-shot, few-shot, and OOD medical image segmentation. It is introduced in arXiv:2607.27856: Benchmarking Foundation and Large Language Models for Few-Shot Medical Image Segmentation. If you use this benchmark, please cite the FAME paper. The BibTeX entry is provided in the Citation section. This Hugging Face dataset contains the released FAME JSON/JSONL protocols plus all referenced image and mask pixels. To keep the repository usable on… See the full description on the dataset page: https://huggingface.co/datasets/Kimokcheon/FAME_benchmark.