Glioma grading method based on graph convolutional network

Peiying Guo; Longfei Li; Chong Shang; Meiyun Wang; Yusong Lin

doi:10.1117/12.2639894

19 October 2022 Glioma grading method based on graph convolutional network

Peiying Guo, Longfei Li, Chong Shang, Meiyun Wang, Yusong Lin

Proceedings Volume 12294, 7th International Symposium on Advances in Electrical, Electronics, and Computer Engineering; 122944J (2022) https://doi.org/10.1117/12.2639894
Event: 7th International Symposium on Advances in Electrical, Electronics and Computer Engineering (ISAEECE 2022), 2022, Xishuangbanna, China

Abstract

The potential relationship information between the multiparameter magnetic resonance (MR) images of similar patients is ignored in the glioma grading model. To address this problem, a multiparameter MR image similarity-aware model based on graph convolutional network (MS-GCN) is proposed for glioma grading. After the features are extracted by a convolutional neural network, the features are constructed into a graph. The image features of patients are used as nodes information in the graph, and the similarity between patient features is used as the weights of edges. The graph convolution operation is used for similarity awareness of the graph data, and the model can simultaneously fuse multiparameter MR image information. The model is verified on the public dataset BraTS2020 and the private dataset GliomaHPPH2018. The AUC scores are 0.970 and 0.988, and the accuracies are 94.7% and 93.6%, respectively. Experimental results show that MS-GCN can accurately capture the similarity between multiparameter MR images and improve the classification performance.

Citation Download Citation

Peiying Guo, Longfei Li, Chong Shang, Meiyun Wang, and Yusong Lin "Glioma grading method based on graph convolutional network", Proc. SPIE 12294, 7th International Symposium on Advances in Electrical, Electronics, and Computer Engineering, 122944J (19 October 2022); https://doi.org/10.1117/12.2639894

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available