Deep Multi-Structural Shape Analysis: Application to Neuroanatomy

TL;DR

This paper introduces a deep neural network that directly analyzes neuroanatomical shapes from raw point clouds, enabling disease prediction and brain age estimation without manual feature extraction.

Contribution

It presents a novel multi-structural deep learning approach that operates on raw point clouds and learns optimal shape representations end-to-end.

Findings

Effective prediction of Alzheimer's disease and mild cognitive impairment.

Accurate regression of brain age.

Visualization of key anatomical regions for predictions.

Abstract

We propose a deep neural network for supervised learning on neuroanatomical shapes. The network directly operates on raw point clouds without the need for mesh processing or the identification of point correspondences, as spatial transformer networks map the data to a canonical space. Instead of relying on hand-crafted shape descriptors, an optimal representation is learned in the end-to-end training stage of the network. The proposed network consists of multiple branches, so that features for multiple structures are learned simultaneously. We demonstrate the performance of our method on two applications: (i) the prediction of Alzheimer's disease and mild cognitive impairment and (ii) the regression of the brain age. Finally, we visualize the important parts of the anatomy for the prediction by adapting the occlusion method to point clouds.