← back to paper
arxiv: 2603.19684 · 2 revisions
TSegAgent: Zero-Shot Tooth Segmentation via Geometry-Aware Vision-Language Agents