Tell2Reg introduces a novel approach that leverages large multimodal models to establish spatial correspondence between images using identical language prompts.

• Eliminates the need for conventional displacement fields or transformation parameters
• Utilizes GroundingDINO and SAM to identify corresponding regions across images
• Provides a fully automated, training-free registration algorithm
• Demonstrates particular effectiveness with medical imaging data

This innovation significantly advances medical image registration capabilities, enabling more accurate alignment of patient images over time or across modalities - crucial for diagnosis, treatment planning, and disease monitoring in clinical settings.

Tell2Reg: Establishing spatial correspondence between images by the same language prompts