One-by-One Stainer: A Fast and Hallucination Resilient Domain Adaptation Method for Histopathology

Karel Moens, Jonas De Vylder, Tinne Tuytelaars, Matthew B. Blaschko
Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, PMLR 315:1539-1564, 2026.

Abstract

Histological staining is a crucial step in the analysis of tissue samples, enabling pathologists to identify and diagnose diseases. However, variations in staining protocols and equipment can lead to inconsistencies in image quality. As a result, histological stain normalization remains an active area of research with increasing demand for AI driven diagnosis support. Unlike other domain adaptation problems, in pathology, the consequences of image hallucinations, obscuring or inserting information, are much greater. We propose a method that mitigates the risk of hallucinations by simplifying the network architecture and training process. Our fully $1\times1$ convolutional architecture prevents textural modifications and we show that a residual connection combined with weight regularization effectively suppresses color information loss. The model does not need supervision from CycleGAN based models. It is trained directly and leverages target domain color distribution information for better convergence without requiring any paired images. As a result, our method can be trained concurrently on images at varying scales, showing differing anatomical structures or dyes. This simplifies the dataset collection, facilitates the adoption at new centers, and reduces the number of models needed at inference.

Cite this Paper

BibTeX
@InProceedings{pmlr-v315-moens26a, title = {One-by-One Stainer: A Fast and Hallucination Resilient Domain Adaptation Method for Histopathology}, author = {Moens, Karel and De Vylder, Jonas and Tuytelaars, Tinne and Blaschko, Matthew B.}, booktitle = {Proceedings of The 9th International Conference on Medical Imaging with Deep Learning}, pages = {1539--1564}, 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/moens26a/moens26a.pdf}, url = {https://proceedings.mlr.press/v315/moens26a.html}, abstract = {Histological staining is a crucial step in the analysis of tissue samples, enabling pathologists to identify and diagnose diseases. However, variations in staining protocols and equipment can lead to inconsistencies in image quality. As a result, histological stain normalization remains an active area of research with increasing demand for AI driven diagnosis support. Unlike other domain adaptation problems, in pathology, the consequences of image hallucinations, obscuring or inserting information, are much greater. We propose a method that mitigates the risk of hallucinations by simplifying the network architecture and training process. Our fully $1\times1$ convolutional architecture prevents textural modifications and we show that a residual connection combined with weight regularization effectively suppresses color information loss. The model does not need supervision from CycleGAN based models. It is trained directly and leverages target domain color distribution information for better convergence without requiring any paired images. As a result, our method can be trained concurrently on images at varying scales, showing differing anatomical structures or dyes. This simplifies the dataset collection, facilitates the adoption at new centers, and reduces the number of models needed at inference.} }
Endnote
%0 Conference Paper %T One-by-One Stainer: A Fast and Hallucination Resilient Domain Adaptation Method for Histopathology %A Karel Moens %A Jonas De Vylder %A Tinne Tuytelaars %A Matthew B. Blaschko %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-moens26a %I PMLR %P 1539--1564 %U https://proceedings.mlr.press/v315/moens26a.html %V 315 %X Histological staining is a crucial step in the analysis of tissue samples, enabling pathologists to identify and diagnose diseases. However, variations in staining protocols and equipment can lead to inconsistencies in image quality. As a result, histological stain normalization remains an active area of research with increasing demand for AI driven diagnosis support. Unlike other domain adaptation problems, in pathology, the consequences of image hallucinations, obscuring or inserting information, are much greater. We propose a method that mitigates the risk of hallucinations by simplifying the network architecture and training process. Our fully $1\times1$ convolutional architecture prevents textural modifications and we show that a residual connection combined with weight regularization effectively suppresses color information loss. The model does not need supervision from CycleGAN based models. It is trained directly and leverages target domain color distribution information for better convergence without requiring any paired images. As a result, our method can be trained concurrently on images at varying scales, showing differing anatomical structures or dyes. This simplifies the dataset collection, facilitates the adoption at new centers, and reduces the number of models needed at inference.
APA
Moens, K., De Vylder, J., Tuytelaars, T. & Blaschko, M.B.. (2026). One-by-One Stainer: A Fast and Hallucination Resilient Domain Adaptation Method for Histopathology. Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 315:1539-1564 Available from https://proceedings.mlr.press/v315/moens26a.html.

Related Material