Joint Learning of Brain Lesion and Anatomy Segmentation from Heterogeneous Datasets
This addresses a practical challenge in neuroimaging for researchers and clinicians by enabling efficient use of publicly available datasets without requiring joint annotations.
The paper tackled the problem of training a single convolutional neural network to simultaneously segment brain lesions and anatomy from disjoint, heterogeneously labeled datasets, and showed that their proposed adaptive cross entropy loss enables such training where standard losses fail, achieving competitive results compared to multi-network approaches.
Brain lesion and anatomy segmentation in magnetic resonance images are fundamental tasks in neuroimaging research and clinical practice. Given enough training data, convolutional neuronal networks (CNN) proved to outperform all existent techniques in both tasks independently. However, to date, little work has been done regarding simultaneous learning of brain lesion and anatomy segmentation from disjoint datasets. In this work we focus on training a single CNN model to predict brain tissue and lesion segmentations using heterogeneous datasets labeled independently, according to only one of these tasks (a common scenario when using publicly available datasets). We show that label contradiction issues can arise in this case, and propose a novel adaptive cross entropy (ACE) loss function that makes such training possible. We provide quantitative evaluation in two different scenarios, benchmarking the proposed method in comparison with a multi-network approach. Our experiments suggest that ACE loss enables training of single models when standard cross entropy and Dice loss functions tend to fail. Moreover, we show that it is possible to achieve competitive results when comparing with multiple networks trained for independent tasks.