I am Junjie Yang, a researcher working on multimodal medical AI, medical image understanding, and open-ended visual generation. My current work focuses on building image-grounded benchmarks that measure whether models can follow clinically meaningful instructions and produce faithful text or image outputs.
My research interests include medical vision-language models, medical image editing, multimodal generation, 3D medical vision, and evaluation of contextual alignment.
Selected Projects
MedGEN-Bench: A Contextually Entangled Benchmark
MedGEN-Bench is a benchmark for open-ended multimodal medical generation. The reported evaluation snapshot contains 6,422 expert-reviewed image–text pairs spanning six canonical imaging modalities, 15 clinical tasks, and 27 named subtasks, including visual question answering, image editing, and contextual multimodal generation.
Publications
- MedGEN-Bench: Contextually entangled benchmark for open-ended multimodal medical generation, Junjie Yang, Yuhao Yan, Gang Wu, et al. arXiv · Project page
News
- 2026: Integrated the MedGEN-Bench project page and its curated multimodal task gallery into this academic homepage.