unORANIC: Unsupervised Orthogonalization of Anatomy and Image-Characteristic Features

TL;DR

unORANIC is an unsupervised method that orthogonalizes anatomy and image features to improve robustness and generalization in medical image analysis, without requiring labels or paired data.

Contribution

It introduces a novel unsupervised orthogonalization technique that enhances image reconstruction and robustness across diverse modalities and tasks.

Findings

Improves classification accuracy on multiple datasets.

Enhances corruption detection and image revision.

Increases robustness and generalizability in medical imaging.

Abstract

We introduce unORANIC, an unsupervised approach that uses an adapted loss function to drive the orthogonalization of anatomy and image-characteristic features. The method is versatile for diverse modalities and tasks, as it does not require domain knowledge, paired data samples, or labels. During test time unORANIC is applied to potentially corrupted images, orthogonalizing their anatomy and characteristic components, to subsequently reconstruct corruption-free images, showing their domain-invariant anatomy only. This feature orthogonalization further improves generalization and robustness against corruptions. We confirm this qualitatively and quantitatively on 5 distinct datasets by assessing unORANIC's classification accuracy, corruption detection and revision capabilities. Our approach shows promise for enhancing the generalizability and robustness of practical applications in medical image analysis. The source code is available at https://github.com/sdoerrich97/unORANIC.