# Probabilistic Point Cloud Reconstructions for Vertebral Shape Analysis

## TL;DR

This paper introduces a probabilistic auto-encoding network for point clouds that captures shape signatures and detects vertebral fractures as anomalies, achieving over 75% AUC without supervision or intensity features.

## Contribution

It presents a novel auto-encoder architecture with a specialized loss and regularization for probabilistic shape analysis of vertebrae, enabling unsupervised fracture detection.

## Key findings

- Achieved >75% AUC in vertebral fracture detection
- Effectively models data variance on unstructured point clouds
- Detects fractures as anomalies without supervision

## Abstract

We propose an auto-encoding network architecture for point clouds (PC) capable of extracting shape signatures without supervision. Building on this, we (i) design a loss function capable of modelling data variance on PCs which are unstructured, and (ii) regularise the latent space as in a variational auto-encoder, both of which increase the auto-encoders' descriptive capacity while making them probabilistic. Evaluating the reconstruction quality of our architectures, we employ them for detecting vertebral fractures without any supervision. By learning to efficiently reconstruct only healthy vertebrae, fractures are detected as anomalous reconstructions. Evaluating on a dataset containing $\sim$1500 vertebrae, we achieve area-under-ROC curve of $>$75%, without using intensity-based features.