New VQA benchmark evaluates VLMs on 3D medical imaging

Arxiv CS.CV (computer vision)

The Oncology VQA benchmark tests vision-language models on 3D medical images, with tasks derived from clinical reports. It aims to provide a scalable, clinically grounded evaluation for VLMs in oncology.

vision-language-modelsbenchmarkscomputer-vision

Arxiv CS.CV (computer vision)·Bo LiuHanxue GuXiangru LiZheren ZhuJacob EllisonKang WangJanine M. LupoYang YangHui Lin

8 days ago