Test-time neural style transfer for domain harmonization, with application to cone detection in adaptive…

Abstract

Domain shift is a well-recognized challenge in medical image analysis using Deep Learning (DL), where variability in acquisition devices, operators, and patient populations can limit model generalization. Adaptive optics scanning laser ophthalmoscopy (AOSLO) imaging exemplifies this issue, as variations in image appearance across acquisitions can affect downstream quantitative analyses. In this work, we repurpose Neural Style Transfer as a test-time augmentation strategy (TT-NST) to harmonize image appearance between training and target domains without model retraining. Our NST approach is model-agnostic and can be applied to standard DL architectures. We evaluated performance, with and without TT-NST, using a well-established DL cone detector (MultiDimensional Recurrent Neural Network) across three datasets spanning far-domain, near-domain, and in-domain imaging conditions. TT-NST substantially improved mean Dice on the far-domain dataset from 0.43 to 0.72, primarily by increasing recall (0.34 to 0.75) with a limited reduction in precision (0.91 to 0.71). More modest gains were observed on the near-domain dataset, while performance slightly decreased on the in-domain dataset, as expected when test images match the training distribution. Compared with alternative test-time augmentation methods, TT-NST achieved the highest mean Dice on far-domain data. These results indicate that TT-NST is a label-free strategy for domain harmonization with potential to improve robustness across medical image analysis tasks under domain shift.

Use of a Retinal Phantom to Facilitate Adaptive Optics Scanning Light Ophthalmoscopy Imaging in a Multicenter Clinical Trial
Translational Vision Science & Technology26 Oct 2026Peer-reviewedCC BY-NC-ND 4.0Full textMedical imaging & radiologyPMC
Representation Transfer of Foundation Models for Ultra-Widefield Retinal Imaging.
Mingya Alexa Gong, Da Ma, Lovre Antonio Budimir +7
IEEE access : practical innovations, open solutions25 Sep 2026Peer-reviewedCC BYFull textMedical imaging & radiologyEurope PMC
Progress in the application of laser photocoagulation therapy in the treatment of pediatric retinal diseases.
Qing Zhang, Xiaoyan Chai, Di Gong +3
Lasers in medical science30 Oct 2026Peer-reviewedCC BY-NC-NDFull textGeneral medicineEurope PMCPMC
Retinal layer thickness prediction from smartphone fundus imaging for early glaucoma detection
R. Prashanth, C. Rajan, K. Geetha
Scientific Reports26 Oct 2026Peer-reviewedCC BY-NC-ND 4.0Full textMedical imaging & radiologyCrossref
F2D-RetinaNet: a feature-to-decision dual-path network with fractional operators for retinal disease detection and grading
Mukesh Delu, Priyanka Harjule, Rajesh Kumar +2
Scientific Reports17 Oct 2026Peer-reviewedCC BY-NC-ND 4.0Full textMedical imaging & radiologyCrossref

Cite

BibTeX RIS

BibTeX

@article{bardet2026testtime,
  title = {{Test-time neural style transfer for domain harmonization, with application to cone detection in adaptive optics retinal imaging}},
  author = {James Bardet and Alexander Aujesky and Alain Jacot-Guillarmod and Victor Amiot and Jelena Potic and Adham Elwakil and Diego Canton and Thomas J. Wolfensberger and Ciara Bergin and Ilenia Meloni and Chiara M. Eandi and Ferenc Sallo Balazs and Adam M. Dubis and Mattia Tomasoni},
  journal = {Scientific Reports},
  year = {2026},
  month = oct,
  doi = {10.1038/s41598-026-74481-9},
  url = {https://www.nature.com/articles/s41598-026-74481-9},
  note = {CC BY 4.0},
}

RIS

TY  - JOUR
TI  - Test-time neural style transfer for domain harmonization, with application to cone detection in adaptive optics retinal imaging
AU  - James Bardet
AU  - Alexander Aujesky
AU  - Alain Jacot-Guillarmod
AU  - Victor Amiot
AU  - Jelena Potic
AU  - Adham Elwakil
AU  - Diego Canton
AU  - Thomas J. Wolfensberger
AU  - Ciara Bergin
AU  - Ilenia Meloni
AU  - Chiara M. Eandi
AU  - Ferenc Sallo Balazs
AU  - Adam M. Dubis
AU  - Mattia Tomasoni
PY  - 2026
DA  - 2026/10/08
JO  - Scientific Reports
DO  - 10.1038/s41598-026-74481-9
UR  - https://www.nature.com/articles/s41598-026-74481-9
AB  - Domain shift is a well-recognized challenge in medical image analysis using Deep Learning (DL), where variability in acquisition devices, operators, and patient populations can limit model generalization. Adaptive optics scanning laser ophthalmoscopy (AOSLO) imaging exemplifies this issue, as variations in image appearance across acquisitions can affect downstream quantitative analyses. In this work, we repurpose Neural Style Transfer as a test-time augmentation strategy (TT-NST) to harmonize image appearance between training and target domains without model retraining. Our NST approach is model-agnostic and can be applied to standard DL architectures. We evaluated performance, with and without TT-NST, using a well-established DL cone detector (MultiDimensional Recurrent Neural Network) across three datasets spanning far-domain, near-domain, and in-domain imaging conditions. TT-NST substantially improved mean Dice on the far-domain dataset from 0.43 to 0.72, primarily by increasing recall (0.34 to 0.75) with a limited reduction in precision (0.91 to 0.71). More modest gains were observed on the near-domain dataset, while performance slightly decreased on the in-domain dataset, as expected when test images match the training distribution. Compared with alternative test-time augmentation methods, TT-NST achieved the highest mean Dice on far-domain data. These results indicate that TT-NST is a label-free strategy for domain harmonization with potential to improve robustness across medical image analysis tasks under domain shift.
SN  - 2045-2322
ER  -

Abstract

Related

Cite