DAM-AL: Dilated Attention Mechanism with Attention Loss for 3D Infant Brain Image Segmentation

TL;DR

This paper introduces DAM-AL, a novel deep learning model with dilated attention and specialized loss functions for improved 3D infant brain MRI segmentation, addressing low contrast and noise challenges.

Contribution

The paper presents a new model combining dilated attention mechanisms and a novel attention loss tailored for infant brain MRI segmentation.

Findings

DAM-AL outperforms state-of-the-art methods on iSeg 2017 dataset.

The model achieves higher Dice coefficients and lower ASD metrics.

Attention mechanisms improve segmentation accuracy in low-contrast MRI images.

Abstract

While Magnetic Resonance Imaging (MRI) has played an essential role in infant brain analysis, segmenting MRI into a number of tissues such as gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) is crucial and complex due to the extremely low intensity contrast between tissues at around 6-9 months of age as well as amplified noise, myelination, and incomplete volume. In this paper, we tackle those limitations by developing a new deep learning model, named DAM-AL, which contains two main contributions, i.e., dilated attention mechanism and hard-case attention loss. Our DAM-AL network is designed with skip block layers and atrous block convolution. It contains both channel-wise attention at high-level context features and spatial attention at low-level spatial structural features. Our attention loss consists of two terms corresponding to region information and hard samples attention. Our proposed DAM-AL has been evaluated on the infant brain iSeg 2017 dataset and the experiments have been conducted on both validation and testing sets. We have benchmarked DAM-AL on Dice coefficient and ASD metrics and compared it with state-of-the-art methods.