Multimodal Glioma Image Segmentation Using Dual Encoder Structure and Channel Spatial Attention Block.

Run Su,Mingquan Ye,Chuandong Cheng,Deyun Zhang,Jinhuai Liu

doi:10.3389/fnins.2020.586197

Abstract

Multimodal medical images provide significant amounts of complementary semantic information. Therefore, multimodal medical imaging has been widely used in the segmentation of gliomas through computational neural networks. However, inputting images from different sources directly to the network does not achieve the best segmentation effect. This paper describes a convolutional neural network called F-S-Net that fuses the information from multimodal medical images and uses the semantic information contained within these images for glioma segmentation. The architecture of F-S-Net is formed by cascading two sub-networks. The first sub-network projects the multimodal medical images into the same semantic space, which ensures they have the same semantic metric. The second sub-network uses a dual encoder structure (DES) and a channel spatial attention block (CSAB) to extract more detailed information and focus on the lesion area. DES and CSAB are integrated into U-Net architectures. A multimodal glioma dataset collected by Yijishan Hospital of Wannan Medical College is used to train and evaluate the network. F-S-Net is found to achieve a dice coefficient of 0.9052 and Jaccard similarity of 0.8280, outperforming several previous segmentation methods.

Highlights

Gliomas, which arise from the canceration of gliocyte in the brain and myelon, are the most common form of cancer in the skull, accounting for 80% of malignant brain tumors (Ostrom et al, 2014)
We propose F-S-Net, which combines image fusion technology to obtain images with richer semantic information
The clinical image data consist of computed tomography (CT) and T2-weighted magnetic resonance imaging (MRI) scans from glioma patients, of which nine images were acquired from low-grade glioma patients and 17 images were obtained from high-grade glioma patients

Summary

Introduction

Gliomas, which arise from the canceration of gliocyte in the brain and myelon, are the most common form of cancer in the skull, accounting for 80% of malignant brain tumors (Ostrom et al, 2014). The early diagnosis and treatment of gliomas are very important. The presence of gliomas can cause complications such as increased intracranial pressure, brain edema, brain hernia, and psychosis. The size, location, and type of a glioma are determined by segmenting the affected region from other normal brain tissue. Accurate segmentation plays an important role in the diagnosis and treatment of gliomas. The automatic segmentation of gliomas would allow doctors to detect the growth of brain tumors earlier and provide additional information for the generation of treatment plans. The automatic segmentation of gliomas would allow doctors to detect the growth of brain tumors earlier and provide additional information for the generation of treatment plans. Bi et al (2019) believed that

Methods

Results

Discussion

Conclusion