Geometric Deep Learning Techniques for Analyzing Brain 3D Meshes

Mariem Ayad, Akrem Sellami, Imed Riadh Farah, Mauro Dalla Mura

Abstract: Geometric deep learning (GDL) has emerged as a powerful paradigm for analyzing complex data represented in non-Euclidean domains. In the field of neuroimaging, 3D meshes have become a prevalent representation for capturing the intricate structures of the brain. This survey paper provides a comprehensive overview of the recent advancements and techniques in using GDL to analyze brain 3D meshes. We systematically review the state-of-the-art methodologies employed in tasks such as segmentation, and classification of brain meshes. Additionally, we discuss the challenges and opportunities in this rapidly evolving field, including data scarcity, interpretability, and scalability. By synthesizing insights from diverse research efforts, this survey aims to guide researchers and practitioners toward a deeper understanding of the application of GDL in neuroimaging and pave the way for future breakthroughs in brain analysis.

External IDs:dblp:conf/atsip/AyadSFM24