Btrfly Net: Vertebrae Labelling with Energy-based Adversarial Learning of Local Spine Prior

TL;DR

This paper introduces Btrfly Net, a novel neural network architecture that leverages sagittal and coronal spine reformations combined with energy-based adversarial training to improve vertebrae labelling accuracy without post-processing.

Contribution

The work presents a new butterfly-shaped network architecture and an energy-based adversarial training method that encodes local spine structure as an anatomical prior.

Findings

Achieves state-of-the-art performance on a benchmark dataset

Operates without post-processing during inference

Effectively encodes local spine structure as a prior

Abstract

Robust localisation and identification of vertebrae is essential for automated spine analysis. The contribution of this work to the task is two-fold: (1) Inspired by the human expert, we hypothesise that a sagittal and coronal reformation of the spine contain sufficient information for labelling the vertebrae. Thereby, we propose a butterfly-shaped network architecture (termed Btrfly Net) that efficiently combines the information across reformations. (2) Underpinning the Btrfly net, we present an energy-based adversarial training regime that encodes local spine structure as an anatomical prior into the network, thereby enabling it to achieve state-of-art performance in all standard metrics on a benchmark dataset of 302 scans without any post-processing during inference.