Skip to content

Read-only version: For security reasons, the Research Information System (FIS) has been temporarily replaced by this read-only version. All public content and full texts remain available; login, editing, submission and statistics are not available. Questions: fis@uni-bamberg.de

Home › Publications › Stylizing ViT : Anatomy-Preserving Instance Style Transfer for Domain Generalization

Publication

Stylizing ViT : Anatomy-Preserving Instance Style Transfer for Domain Generalization

Doerrich, Sebastian; Di Salvo, Francesco; Alle, Jonas; u. a. (2026): Stylizing ViT : Anatomy-Preserving Instance Style Transfer for Domain Generalization, in: 2026 IEEE 23rd International Symposium on Biomedical Imaging (ISBI), Piscataway, NJ: IEEE, doi: 10.1109/isbi61048.2026.11515498.

Details

Faculty/Chair
Faculty/Professorship
Author
Title of the compilation
2026 IEEE 23rd International Symposium on Biomedical Imaging (ISBI)
Conference
23rd International Symposium on Biomedical Imaging (ISBI), 08-11 April 2026, London
Publisher Information
IEEE : Piscataway, NJ
Year of publication
2026
Pages
5
Language
English
DOI
URL
Abstract

Deep learning models in medical image analysis often struggle with generalizability across domains and demographic groups due to data heterogeneity and scarcity. Traditional augmentation improves robustness, but fails under substantial domain shifts. Recent advances in stylistic augmentation enhance domain generalization by varying image styles but fall short in terms of style diversity or by introducing artifacts into the generated images. To address these limitations, we propose Stylizing ViT, a novel Vision Transformer encoder that utilizes weight-shared attention blocks for both self- and cross-attention. This design allows the same attention block to maintain anatomical consistency through self-attention while performing style transfer via cross-attention. We assess the effectiveness of our method for domain generalization by employing it for data augmentation on three distinct image classification tasks in the context of histopathology and dermatology. Results demonstrate an improved robustness (up to +13% accuracy) over the state of the art while generating perceptually convincing images without artifacts. Additionally, we show that Stylizing ViT is effective beyond training, achieving a 17% performance improvement during inference when used for test-time augmentation. The source code is available at https://github.com/sdoerrich97/stylizing-vit .

Keywords
style transfer
domain generalization
crossattention
data augmentation
vision transformer
Peer Reviewed
yes
International Distribution
yes
Type
Conferenceobject
URI
Activation date
2026-09-09

Permalink: https://fis.uni-bamberg.de/handle/uniba/117126 · Data as JSON · Cite: BibTeX · RIS

Data as of: 2026-11-16

Restricted read-only access: login, editing and submission are temporarily disabled.