LaB-GATr: geometric algebra transformers for large biomedical surface and volume meshes

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Suk, Julian, Imre, Baris, Wolterink, Jelmer M.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913569948303360
author Suk, Julian
Imre, Baris
Wolterink, Jelmer M.
author_facet Suk, Julian
Imre, Baris
Wolterink, Jelmer M.
contents Many anatomical structures can be described by surface or volume meshes. Machine learning is a promising tool to extract information from these 3D models. However, high-fidelity meshes often contain hundreds of thousands of vertices, which creates unique challenges in building deep neural network architectures. Furthermore, patient-specific meshes may not be canonically aligned which limits the generalisation of machine learning algorithms. We propose LaB-GATr, a transfomer neural network with geometric tokenisation that can effectively learn with large-scale (bio-)medical surface and volume meshes through sequence compression and interpolation. Our method extends the recently proposed geometric algebra transformer (GATr) and thus respects all Euclidean symmetries, i.e. rotation, translation and reflection, effectively mitigating the problem of canonical alignment between patients. LaB-GATr achieves state-of-the-art results on three tasks in cardiovascular hemodynamics modelling and neurodevelopmental phenotype prediction, featuring meshes of up to 200,000 vertices. Our results demonstrate that LaB-GATr is a powerful architecture for learning with high-fidelity meshes which has the potential to enable interesting downstream applications. Our implementation is publicly available.
format Preprint
id arxiv_https___arxiv_org_abs_2403_07536
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle LaB-GATr: geometric algebra transformers for large biomedical surface and volume meshes
Suk, Julian
Imre, Baris
Wolterink, Jelmer M.
Computer Vision and Pattern Recognition
Machine Learning
Many anatomical structures can be described by surface or volume meshes. Machine learning is a promising tool to extract information from these 3D models. However, high-fidelity meshes often contain hundreds of thousands of vertices, which creates unique challenges in building deep neural network architectures. Furthermore, patient-specific meshes may not be canonically aligned which limits the generalisation of machine learning algorithms. We propose LaB-GATr, a transfomer neural network with geometric tokenisation that can effectively learn with large-scale (bio-)medical surface and volume meshes through sequence compression and interpolation. Our method extends the recently proposed geometric algebra transformer (GATr) and thus respects all Euclidean symmetries, i.e. rotation, translation and reflection, effectively mitigating the problem of canonical alignment between patients. LaB-GATr achieves state-of-the-art results on three tasks in cardiovascular hemodynamics modelling and neurodevelopmental phenotype prediction, featuring meshes of up to 200,000 vertices. Our results demonstrate that LaB-GATr is a powerful architecture for learning with high-fidelity meshes which has the potential to enable interesting downstream applications. Our implementation is publicly available.
title LaB-GATr: geometric algebra transformers for large biomedical surface and volume meshes
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2403.07536