Caroline Magg won the award for the best short oral presentation at the MIDL Young Researcher Showcase for her presentation titled “Slices of Insight: Zero-shot Bone CT segmentation for Promptable Foundation Models ”!

At the showcase, she presented some first results from her ongoing study with the aim to test the prompt robustness of promptable foundation models for bone CT segmentation. 11 existing models have been identified – with various model sizes and settings – and evaluated automatically from the reference mask generated prompts. The results show some clear trends and strong candidates, which will inform the model selection for the following prompt robustness analysis.