Learning Domain-Agnostic Visual Representation for Computational Pathology Using Medically-Irrelevant Style Transfer Augmentation

• Published in: IEEE transactions on medical imaging Vol. 40; no. 12; pp. 3945 - 3954
• Main Authors: Yamashita, Rikiya, Long, Jin, Banda, Snikitha, Shen, Jeanne, Rubin, Daniel L.
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.12.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Biomedical imaging; Computational modeling; Computational pathology; Computer applications; Data augmentation; Data models; Deep learning; domain-agnostic learning; Domains; Histopathology; Learning algorithms; Machine Learning; Medical imaging; Pathology; Representations; Semantics; style transfer; Task analysis; Visual discrimination learning; Visualization
• ISSN: 0278-0062; 1558-254X; 1558-254X
• DOI: 10.1109/TMI.2021.3101985

Suboptimal generalization of machine learning models on unseen data is a key challenge which hampers the clinical applicability of such models to medical imaging. Although various methods such as domain adaptation and domain generalization have evolved to combat this challenge, learning robust and generalizable representations is core to medical image understanding, and continues to be a problem. Here, we propose STRAP ( S tyle TR ansfer A ugmentation for histo P athology), a form of data augmentation based on random style transfer from non-medical style sources such as artistic paintings, for learning domain-agnostic visual representations in computational pathology. Style transfer replaces the low-level texture content of an image with the uninformative style of randomly selected style source image, while preserving the original high-level semantic content. This improves robustness to domain shift and can be used as a simple yet powerful tool for learning domain-agnostic representations. We demonstrate that STRAP leads to state-of-the-art performance, particularly in the presence of domain shifts, on two particular classification tasks in computational pathology. Our code is available at https://github.com/rikiyay/style-transfer-for-digital-pathology .