Y-net: 3D intracranial artery segmentation using a convolutional autoencoder

TL;DR

This paper introduces Y-net, a convolutional autoencoder designed for 3D intracranial artery segmentation on MRA images, demonstrating superior performance over traditional methods.

Contribution

The paper presents an optimized CAE model, Y-net, specifically tailored for 3D intracranial artery segmentation, improving accuracy and noise reduction.

Findings

Y-net outperforms traditional segmentation methods in accuracy.

Y-net effectively reduces noise and extracts features from MRA data.

The model was trained on 49 cases, showing promising results.

Abstract

Automated segmentation of intracranial arteries on magnetic resonance angiography (MRA) allows for quantification of cerebrovascular features, which provides tools for understanding aging and pathophysiological adaptations of the cerebrovascular system. Using a convolutional autoencoder (CAE) for segmentation is promising as it takes advantage of the autoencoder structure in effective noise reduction and feature extraction by representing high dimensional information with low dimensional latent variables. In this report, an optimized CAE model (Y-net) was trained to learn a 3D segmentation model of intracranial arteries from 49 cases of MRA data. The trained model was shown to perform better than the three traditional segmentation methods in both binary classification and visual evaluation.