Skeletal Point Representations with Geometric Deep Learning
This work addresses the need for efficient skeletal modeling in medical imaging, though it appears incremental as it builds on existing learning-based methods with added geometric losses.
The paper tackles the problem of time-consuming manual skeletonization of anatomical structures by proposing novel geometric terms for learning-based skeletal extraction, resulting in similar accuracy to traditional methods but with significantly faster computation.
Skeletonization has been a popular shape analysis technique that models both the interior and exterior of an object. Existing template-based calculations of skeletal models from anatomical structures are a time-consuming manual process. Recently, learning-based methods have been used to extract skeletons from 3D shapes. In this work, we propose novel additional geometric terms for calculating skeletal structures of objects. The results are similar to traditional fitted s-reps but but are produced much more quickly. Evaluation on real clinical data shows that the learned model predicts accurate skeletal representations and shows the impact of proposed geometric losses along with using s-reps as weak supervision.