Search papers, labs, and topics across Lattice.
This paper introduces Confidence-Guided Diffusion Refinement (CoDiR), a novel semi-supervised framework for histopathology segmentation that enhances pseudo-labels in ambiguous regions using a conditional diffusion model. By integrating a Mean Teacher model with a focus on low-confidence areas, CoDiR achieves significant improvements in segmentation accuracy, reaching up to 90.29% mDice on the CRAG dataset with only 20% labeled data. The results demonstrate that the refinement module is crucial, contributing an additional 6.36% mDice over the baseline, thus setting a new benchmark in the field.
Refining low-confidence pseudo-labels with diffusion techniques boosts histopathology segmentation accuracy by over 6% compared to traditional methods.
Semi-supervised histopathology segmentation is challenging due to scarce annotations and unreliable pseudo-labels in ambiguous gland regions. To address this problem, we propose Confidence-Guided Diffusion Refinement (CoDiR), a semi-supervised framework that combines a Mean Teacher segmentation model with diffusion-based pseudo-label refinement. Given an unlabeled image, the teacher first produces a soft prediction, and only low-confidence regions are refined by a conditional diffusion model trained to capture plausible mask structures from labeled data. The refined mask is then fused with reliable teacher predictions and used to train the student with confidence weighting and consistency regularization. On the GlaS and CRAG datasets CoDiR reaches 88.09\% and 89.83\% mDice with 10\% labeled data, and 89.19\% and 90.29\% mDice with 20\%, matching or exceeding the strongest published method on seven of the eight benchmark metrics. Ablations attribute the largest single contribution to the refinement module, which adds +6.36\% mDice over the Mean Teacher baseline. The implementation code is publicly available at: https://github.com/vongla345/codir
