RNCA: Self-Repairing Segmentation Masks

Malte Silbernagel, Albert Alonso, Jens Petersen, Bulat Ibragimov, Marleen de Bruijne, Madeleine K. Wyburd
Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, PMLR 315:2474-2495, 2026.

Abstract

Accurately predicting topologically correct masks remains a difficult task for general segmentation models, which often produce fragmented or disconnected outputs. Fixing these artifacts typically requires handcrafted refinement rules or architectures specialized to a particular task. Here, we show that Neural Cellular Automata (NCA) can be directly repurposed as an effective refinement mechanism, using local, iterative updates guided by image context to repair segmentation masks. By training on imperfect masks and ground truths, the automaton learns the structural properties of the target shape while relying solely on local information. When applied to coarse, globally predicted masks, the learned dynamics progressively reconnect broken regions, prune loose fragments and converge towards stable, topologically consistent results. We show how refinement NCA () can be easily applied to repair common topological errors produced by different base segmentation models and tasks: for fragmented retinal vessels, it yields 2–3% gains in Dice/clDice and improves Betti Errors, reducing $\beta_0$ errors by 60% and $\beta_1$ by 20%; for myocardium, it repairs 61.5% of broken cases in a zero-shot setting while lowering ASSD and HD by 19% and 16%, respectively. This showcases NCA as effective and broadly applicable refiners.

Cite this Paper

BibTeX
@InProceedings{pmlr-v315-silbernagel26a, title = {RNCA: Self-Repairing Segmentation Masks}, author = {Silbernagel, Malte and Alonso, Albert and Petersen, Jens and Ibragimov, Bulat and de Bruijne, Marleen and Wyburd, Madeleine K.}, booktitle = {Proceedings of The 9th International Conference on Medical Imaging with Deep Learning}, pages = {2474--2495}, year = {2026}, editor = {Huo, Yuankai and Gao, Mingchen and Kuo, Chang-Fu and Jin, Yueming and Deng, Ruining}, volume = {315}, series = {Proceedings of Machine Learning Research}, month = {08--10 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v315/main/assets/silbernagel26a/silbernagel26a.pdf}, url = {https://proceedings.mlr.press/v315/silbernagel26a.html}, abstract = {Accurately predicting topologically correct masks remains a difficult task for general segmentation models, which often produce fragmented or disconnected outputs. Fixing these artifacts typically requires handcrafted refinement rules or architectures specialized to a particular task. Here, we show that Neural Cellular Automata (NCA) can be directly repurposed as an effective refinement mechanism, using local, iterative updates guided by image context to repair segmentation masks. By training on imperfect masks and ground truths, the automaton learns the structural properties of the target shape while relying solely on local information. When applied to coarse, globally predicted masks, the learned dynamics progressively reconnect broken regions, prune loose fragments and converge towards stable, topologically consistent results. We show how refinement NCA () can be easily applied to repair common topological errors produced by different base segmentation models and tasks: for fragmented retinal vessels, it yields 2–3% gains in Dice/clDice and improves Betti Errors, reducing $\beta_0$ errors by 60% and $\beta_1$ by 20%; for myocardium, it repairs 61.5% of broken cases in a zero-shot setting while lowering ASSD and HD by 19% and 16%, respectively. This showcases NCA as effective and broadly applicable refiners.} }
Endnote
%0 Conference Paper %T RNCA: Self-Repairing Segmentation Masks %A Malte Silbernagel %A Albert Alonso %A Jens Petersen %A Bulat Ibragimov %A Marleen de Bruijne %A Madeleine K. Wyburd %B Proceedings of The 9th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Yuankai Huo %E Mingchen Gao %E Chang-Fu Kuo %E Yueming Jin %E Ruining Deng %F pmlr-v315-silbernagel26a %I PMLR %P 2474--2495 %U https://proceedings.mlr.press/v315/silbernagel26a.html %V 315 %X Accurately predicting topologically correct masks remains a difficult task for general segmentation models, which often produce fragmented or disconnected outputs. Fixing these artifacts typically requires handcrafted refinement rules or architectures specialized to a particular task. Here, we show that Neural Cellular Automata (NCA) can be directly repurposed as an effective refinement mechanism, using local, iterative updates guided by image context to repair segmentation masks. By training on imperfect masks and ground truths, the automaton learns the structural properties of the target shape while relying solely on local information. When applied to coarse, globally predicted masks, the learned dynamics progressively reconnect broken regions, prune loose fragments and converge towards stable, topologically consistent results. We show how refinement NCA () can be easily applied to repair common topological errors produced by different base segmentation models and tasks: for fragmented retinal vessels, it yields 2–3% gains in Dice/clDice and improves Betti Errors, reducing $\beta_0$ errors by 60% and $\beta_1$ by 20%; for myocardium, it repairs 61.5% of broken cases in a zero-shot setting while lowering ASSD and HD by 19% and 16%, respectively. This showcases NCA as effective and broadly applicable refiners.
APA
Silbernagel, M., Alonso, A., Petersen, J., Ibragimov, B., de Bruijne, M. & Wyburd, M.K.. (2026). RNCA: Self-Repairing Segmentation Masks. Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 315:2474-2495 Available from https://proceedings.mlr.press/v315/silbernagel26a.html.

Related Material