Convolutional neural networks (CNNs) allow for parameter sharing and translational equivariance by using convolutional kernels in their linear layers. By restricting these kernels to be SO(3)-steerable, CNNs can further improve parameter sharing and equivariance. These equivariant convolutional layers have several advantages over standard convolutional layers, including increased robustness to unseen poses, smaller network size, and improved sample efficiency. Despite this, most segmentation networks used in medical image analysis continue to rely on standard convolutional kernels. In this paper, we present a new family of segmentation networks that use equivariant voxel convolutions based on spherical harmonics. These SE(3)-equivariant volumetric segmentation networks, which are robust to data poses not seen during training, do not require rotation-based data augmentation during training. In addition, we demonstrate improved segmentation performance in MRI brain tumor and healthy brain structure segmentation tasks, with enhanced robustness to reduced amounts of training data and improved parameter efficiency. Code to reproduce our results, and to implement the equivariant segmentation networks for other tasks is available at http://github.com/SCAN-NRAD/e3nn_Unet.
MRI · segmentation · rotation equivariance
@article{melba:2024:010:diaz,
title = "Leveraging SO(3)-steerable convolutions for pose-robust semantic segmentation in 3D medical data",
author = "Diaz, Ivan and Geiger, Mario and McKinley, Richard Iain",
journal = "Machine Learning for Biomedical Imaging",
volume = "2",
issue = "May 2024 issue",
year = "2024",
pages = "834--855",
issn = "2766-905X",
doi = "https://doi.org/10.59275/j.melba.2024-7189",
url = "https://melba-journal.org/2024:010"
}
TY - JOUR
AU - Diaz, Ivan
AU - Geiger, Mario
AU - McKinley, Richard Iain
PY - 2024
TI - Leveraging SO(3)-steerable convolutions for pose-robust semantic segmentation in 3D medical data
T2 - Machine Learning for Biomedical Imaging
VL - 2
IS - May 2024 issue
SP - 834
EP - 855
SN - 2766-905X
DO - https://doi.org/10.59275/j.melba.2024-7189
UR - https://melba-journal.org/2024:010
ER -