U-SegNet: Fully Convolutional Neural Network based Automated Brain tissue segmentation Tool

TL;DR

This paper introduces U-SegNet, a hybrid deep learning model combining SegNet and U-Net architectures, to improve brain tissue segmentation accuracy in MRI images, aiding neuro-disorder diagnosis.

Contribution

The paper proposes a novel FCN architecture with skip connections inspired by U-Net integrated into SegNet for enhanced tissue boundary detection.

Findings

Achieved an average dice ratio of 89.74% on IBSR dataset.

Demonstrated improved segmentation performance over existing methods.

Effectively handled thin and smooth tissue boundaries in MRI images.

Abstract

Automated brain tissue segmentation into white matter (WM), gray matter (GM), and cerebro-spinal fluid (CSF) from magnetic resonance images (MRI) is helpful in the diagnosis of neuro-disorders such as epilepsy, Alzheimer's, multiple sclerosis, etc. However, thin GM structures at the periphery of cortex and smooth transitions on tissue boundaries such as between GM and WM, or WM and CSF pose difficulty in building a reliable segmentation tool. This paper proposes a Fully Convolutional Neural Network (FCN) tool, that is a hybrid of two widely used deep learning segmentation architectures SegNet and U-Net, for improved brain tissue segmentation. We propose a skip connection inspired from U-Net, in the SegNet architetcure, to incorporate fine multiscale information for better tissue boundary identification. We show that the proposed U-SegNet architecture, improves segmentation performance, as measured by average dice ratio, to 89.74% on the widely used IBSR dataset consisting of T-1 weighted MRI volumes of 18 subjects.