A Lifelong Learning Approach to Brain MR Segmentation Across Scanners and Protocols

TL;DR

This paper presents a lifelong multi-domain learning method for brain MRI segmentation that adapts to different scanners and protocols using a single CNN with domain-specific batch normalization, requiring minimal labeled data.

Contribution

The authors introduce a CNN architecture with shared filters and domain-specific batch normalization layers that can quickly adapt to new MRI domains with few labeled images, while maintaining performance on previous domains.

Findings

Significantly reduces domain adaptation effort in MRI segmentation.

Achieves performance close to dedicated models for each scanner.

Requires only about 4 labeled images for new domains.

Abstract

Convolutional neural networks (CNNs) have shown promising results on several segmentation tasks in magnetic resonance (MR) images. However, the accuracy of CNNs may degrade severely when segmenting images acquired with different scanners and/or protocols as compared to the training data, thus limiting their practical utility. We address this shortcoming in a lifelong multi-domain learning setting by treating images acquired with different scanners or protocols as samples from different, but related domains. Our solution is a single CNN with shared convolutional filters and domain-specific batch normalization layers, which can be tuned to new domains with only a few ($\approx$ 4) labelled images. Importantly, this is achieved while retaining performance on the older domains whose training data may no longer be available. We evaluate the method for brain structure segmentation in MR images. Results demonstrate that the proposed method largely closes the gap to the benchmark, which is training a dedicated CNN for each scanner.