A Lifelong Learning Approach to Brain MR Segmentation Across Scanners and Protocols
This addresses the practical limitation of CNNs in medical imaging by enabling adaptation to new scanners/protocols with minimal data, though it is incremental as it builds on existing multi-domain learning techniques.
The paper tackles the problem of CNN performance degradation in brain MR segmentation when applied to images from different scanners or protocols, proposing a lifelong learning approach that uses a single CNN with shared filters and domain-specific batch normalization, achieving results close to training dedicated CNNs per scanner with only about 4 labeled images per new domain.
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.