Learning Shape Representation on Sparse Point Clouds for Volumetric Image Segmentation
This work addresses medical image segmentation problems for clinicians, but it is incremental as it builds upon existing PointCNN methods.
The paper tackles the challenges of volumetric image segmentation in medical images, such as large volumes and class imbalances, by proposing point-wise classification of sparse point clouds to jointly encode shape and volumetric information, which yields significantly improved results in peripheral nerve segmentation from MRI images.
Volumetric image segmentation with convolutional neural networks (CNNs) encounters several challenges, which are specific to medical images. Among these challenges are large volumes of interest, high class imbalances, and difficulties in learning shape representations. To tackle these challenges, we propose to improve over traditional CNN-based volumetric image segmentation through point-wise classification of point clouds. The sparsity of point clouds allows processing of entire image volumes, balancing highly imbalanced segmentation problems, and explicitly learning an anatomical shape. We build upon PointCNN, a neural network proposed to process point clouds, and propose here to jointly encode shape and volumetric information within the point cloud in a compact and computationally effective manner. We demonstrate how this approach can then be used to refine CNN-based segmentation, which yields significantly improved results in our experiments on the difficult task of peripheral nerve segmentation from magnetic resonance neurography images. By synthetic experiments, we further show the capability of our approach in learning an explicit anatomical shape representation.