Reliable Source Approximation: Source-Free Unsupervised Domain Adaptation for Vestibular Schwannoma MRI Segmentation

TL;DR

This paper introduces Reliable Source Approximation (RSA), a novel source-free unsupervised domain adaptation method using diffusion models and uncertainty estimation to improve MRI segmentation of vestibular schwannoma across different MRI sequences.

Contribution

RSA is the first approach to generate source-like images from target MRI data using a conditional diffusion model for domain adaptation in medical image segmentation.

Findings

RSA outperforms existing SFUDA methods in vestibular schwannoma segmentation.

The method effectively generates reliable pseudo labels for target domain images.

Experimental results show consistent improvement across multi-modality MRI datasets.

Abstract

Source-Free Unsupervised Domain Adaptation (SFUDA) has recently become a focus in the medical image domain adaptation, as it only utilizes the source model and does not require annotated target data. However, current SFUDA approaches cannot tackle the complex segmentation task across different MRI sequences, such as the vestibular schwannoma segmentation. To address this problem, we proposed Reliable Source Approximation (RSA), which can generate source-like and structure-preserved images from the target domain for updating model parameters and adapting domain shifts. Specifically, RSA deploys a conditional diffusion model to generate multiple source-like images under the guidance of varying edges of one target image. An uncertainty estimation module is then introduced to predict and refine reliable pseudo labels of generated images, and the prediction consistency is developed to select the most reliable generations. Subsequently, all reliable generated images and their pseudo labels are utilized to update the model. Our RSA is validated on vestibular schwannoma segmentation across multi-modality MRI. The experimental results demonstrate that RSA consistently improves domain adaptation performance over other state-of-the-art SFUDA methods. Code is available at https://github.com/zenghy96/Reliable-Source-Approximation.