The Role Of Recurrent Refinement In Medical Image Segmentation
Summary AI-generated
- TL;DR
- While feed-forward segmentation networks perform well on large cohorts, they degrade in constrained clinical settings with few training samples and imaging artifacts.
- Problem
- Human vision, by contrast, remains robust under these conditions, a property partly attributed to the recurrent and feedback connectivity of the visual cortex, which refines visual information over time.
- Method
- Not specified in the abstract.
- Results
- Evaluated on two open-source datasets, Kvasir- SEG and ARCADE, our recurrent blocks add less than 1.5% parameters and yield gains on clean data (up to 1.4 pp) and under noise (up to 17 pp) while reducing HD95.
- Contributions
- Not specified in the abstract.
- Limitations
- Not specified in the abstract.
- Takeaways
- Code is available at https://github.com/joaodsmarques/recurrencehelpsegmentation .
- Applications
- Not specified in the abstract.
- Topics
- arXiv (Cornell University); The Role Of Recurrent Refinement In Medical Image Segmentation
- For industry
- Potentially relevant to healthcare and clinical decision support.
- Why it matters
- Advances methods that could benefit healthcare and clinical decision support.
Abstract
While feed-forward segmentation networks perform well on large cohorts, they degrade in constrained clinical settings with few training samples and imaging artifacts. Human vision, by contrast, remains robust under these conditions, a property partly attributed to the recurrent and feedback connectivity of the visual cortex, which refines visual information over time. Inspired by this, we append lightweight recurrent refinement blocks to segmentation backbones, namely Conditional Random Fields (CRFs), modern Hopfield networks and convolutional GRUs, and study their behaviour as the number of iterations increases at test time (including beyond the training horizon). Evaluated on two open-source datasets, Kvasir- SEG and ARCADE, our recurrent blocks add less than 1.5% parameters and yield gains on clean data (up to 1.4 pp) and under noise (up to 17 pp) while reducing HD95. Code is available at https://github.com/joaodsmarques/recurrencehelpsegmentation .