Authors: Xiang Chen, Min Liu, Rongguang Wang, Renjiu Hu, Dongdong Liu, Gaolei Li, Hang Zhang

Published on: November 27, 2023

Impact Score: 8.15

Arxiv code: Arxiv:2311.15607

Summary

• What is new: Introduces textSCF, a method that combines spatially covariant filters with textual anatomical prompts for image registration, breaking conventional translation-invariance constraints.
• Why this is important: Existing methods for medical image registration lack efficiency and fail to leverage anatomical priors effectively.
• What the research proposes: textSCF utilizes visual-language models to encode anatomical prompts, optimizing an implicit function for better registration accuracy and efficiency.
• Results: Outperformed state-of-the-art models in the MICCAI Learn2Reg 2021 challenge, especially in brain MRI and abdominal CT tasks, showing significant improvements in computational efficiency and accuracy.

Technical Details

Technological frameworks used: nan

Models used: Spatially Covariant Filters, Visual-Language Models

Data used: Inter-subject brain MRI and abdominal CT images

Potential Impact

Healthcare imaging companies, clinical diagnostic tools, medical AI software developers

Want to implement this idea in a business?

We have generated a startup concept here: Anatomee.