3D Segmentation Networks for Excessive Numbers of Classes: Distinct Bone Segmentation in Upper Bodies
This addresses a domain-specific problem for medical imaging, enabling detailed bone segmentation for diagnosis and surgical planning, but it is incremental as it adapts existing methods to handle many classes.
The paper tackles the problem of segmenting 125 distinct bones in 3D CT scans, which is more classes than typical 3D segmentation tasks, and demonstrates a method that automatically segments over one hundred bones simultaneously in an end-to-end fashion.
Segmentation of distinct bones plays a crucial role in diagnosis, planning, navigation, and the assessment of bone metastasis. It supplies semantic knowledge to visualisation tools for the planning of surgical interventions and the education of health professionals. Fully supervised segmentation of 3D data using Deep Learning methods has been extensively studied for many tasks but is usually restricted to distinguishing only a handful of classes. With 125 distinct bones, our case includes many more labels than typical 3D segmentation tasks. For this reason, the direct adaptation of most established methods is not possible. This paper discusses the intricacies of training a 3D segmentation network in a many-label setting and shows necessary modifications in network architecture, loss function, and data augmentation. As a result, we demonstrate the robustness of our method by automatically segmenting over one hundred distinct bones simultaneously in an end-to-end learnt fashion from a CT-scan.