Attention Gate ResU-Net for Automatic MRI Brain Tumor Segmentation

Abstract

Brain tumor segmentation technology plays a pivotal role in the process of diagnosis and treatment of MRI brain tumors. It helps doctors to locate and measure tumors, as well as develop treatment and rehabilitation strategies. Recently, MRI brain tumor segmentation methods based on U-Net architecture have become popular as they largely improve the segmentation accuracy by applying skip connection to combine high-level feature information and low-level feature information. Meanwhile, researchers have demonstrated that introducing attention mechanism into U-Net can enhance local feature expression and improve the performance of medical image segmentation. In this work, we aim to explore the effectiveness of a recent attention module called attention gate for brain tumor segmentation task, and a novel Attention Gate Residual U-Net model, i.e., AGResU-Net, is further presented. AGResU-Net integrates residual modules and attention gates with a primeval and single U-Net architecture, in which a series of attention gate units are added into the skip connection for highlighting salient feature information while disambiguating irrelevant and noisy feature responses. AGResU-Net not only extracts abundant semantic information to enhance the ability of feature learning, but also pays attention to the information of small-scale brain tumors. We extensively evaluate attention gate units on three authoritative MRI brain tumor benchmarks, i.e., BraTS 2017, BraTS 2018 and BraTS 2019. Experimental results illuminate that models with attention gate units, i.e., Attention Gate U-Net (AGU-Net) and AGResU-Net, outperform their baselines of U-Net and ResU-Net, respectively. In addition, AGResU-Net achieves competitive performance than the representative brain tumor segmentation methods.

Dual‐Path Attention Compensation U‐Net for Stroke Lesion Segmentation

“…The DPAC structure proposed in this paper that uses the auxiliary network to compensate the primary network can be applied to most segmentation models with spatial self-attention. We implemented our method on two other segmentation models with self-attention mechanism, RA-UNet [ 30 ] and AGResU-Net [ 31 ], and compared the experimental results of single-path with dual-path networks with auxiliary networks. The experimental results are shown in Table 7 .…”

Brain Image Segmentation in Recent Years: A Narrative Review

“…Deep learning using 2D images requires brain image slices or extracted 2D patches from 3D images as an input for the 2D convolutional kernel. Several studies [21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37] have been published on the deep learning-based method using 2D images. Sergio Pereira et al [21] introduced cascade layers using small 3*3 convolutions kernels to reduce overfitting.…”

“…In addition to the various methods proposed, Zhang et al [31] presented a residual U-Net and attention mechanism in a unified architecture named AGResU-Net for patch-wise brain tumor segmentation. Attention gate units were added into the up-skip connection of the U-Net structure to highlight the important feature details along with disambiguates in noise and irrelevant feature responses.…”

Brain tumor detection and classification using machine learning: a comprehensive survey

“…The classifiers are fed the salient features derived from its pictures. This method has been tested on BRATS 2017, 2018 and 2019 datasets [244]. The tumor region is localized on Flair sequences of brats 2012 series.…”