19 May 2020
Congratulations to Leonie Henschel for the 1st place: Best Scientific Submission at the 2020 BVM Workshop. Our submission Parameter Space CNN for Cortical Surface Segmentation and YouTube presentation has convinced the reviewers.
In that work we investigate neural networks for spherical signals of the brain:
Spherical coordinate systems have become a standard for analyzing human cortical neuroimaging data. Surface-based signals, such as curvature, folding patterns, functional activations, or estimates of myelination define relevant cortical regions. Surface-based deep learning approaches, however, such as spherical CNNs primarily focus on classification and cannot yet achieve satisfactory accuracy in segmentation tasks. To perform surface-based segmentation of the human cortex, we introduce and evaluate a 2D parameter space approach with view aggregation (p3CNN). We evaluate this network with respect to accuracy and show that it outperforms the spherical CNN by a margin, increasing the average Dice similarity score for cortical segmentation to above 0.9.
Figure: Comparison of a spherical CNN (ugscnn, middle) and our proposed view-aggregation on 2D spherical parameter spaces (p3CNN, right) for cortical segmentation of the cortex (bottom row).