Unsupervised Feature Orthogonalization for Learning Distortion-Invariant Representations
This method addresses the need for advanced medical image analysis in resource-constrained environments lacking large datasets, though it appears incremental as it builds on existing techniques like feature orthogonalization and Vision Transformers.
The paper tackles the problem of learning robust and unbiased representations for medical image analysis by introducing unORANIC+, which integrates unsupervised feature orthogonalization with Vision Transformers to separate anatomical and image-specific attributes, resulting in excellent performance across various tasks and datasets.
This study introduces unORANIC+, a novel method that integrates unsupervised feature orthogonalization with the ability of a Vision Transformer to capture both local and global relationships for improved robustness and generalizability. The streamlined architecture of unORANIC+ effectively separates anatomical and image-specific attributes, resulting in robust and unbiased latent representations that allow the model to demonstrate excellent performance across various medical image analysis tasks and diverse datasets. Extensive experimentation demonstrates unORANIC+'s reconstruction proficiency, corruption resilience, as well as capability to revise existing image distortions. Additionally, the model exhibits notable aptitude in downstream tasks such as disease classification and corruption detection. We confirm its adaptability to diverse datasets of varying image sources and sample sizes which positions the method as a promising algorithm for advanced medical image analysis, particularly in resource-constrained environments lacking large, tailored datasets. The source code is available at https://github.com/sdoerrich97/unoranic-plus .