Title:
DISEASE PREDICTION SYSTEM USING GRAPH CONVOLUTIONAL NEURAL NETWORK AND BASED ON MULTI-MODAL MAGNETIC RESONANCE IMAGING
Document Type and Number:
WIPO Patent Application WO/2024/083057
Kind Code:
A1
Abstract:
Disclosed in the present invention is a disease prediction system using a graph convolutional neural network and based on multi-modal magnetic resonance imaging. Radiomics information of a plurality of cross-modal encephalic regions is extracted from multi-modal magnetic resonance data and is taken as features of nodes, and connectomics information between the encephalic regions is extracted to form an adjacent matrix. A T1 weighted structure image is subjected to cortex information extraction by means of cortex reconstruction, and resting-state magnetic resonance data is used for calculating an amplitude of low frequency fluctuations, and regional homogeneity and functional connectivity. By means of multi-modal data preprocessing, image indicator extraction and structured data integration, multi-modal unstructured magnetic resonance image data is integrated into unified graph structure data, and diseases are predicted by using a method based on a graph convolutional neural network. A multi-modal feature obtained by means of calculation in the present invention is used for performing disease prediction, such that cross-modal physiological indicators of a plurality of encephalic regions and the correlation between the encephalic regions can be better combined, and the prediction capability of a model and the generalization capability of the model with respect to different diseases are improved.
Inventors:
ZHANG YU (CN)
SUN CHAOLIANG (CN)
WANG ZHICHAO (CN)
ZHANG HUAN (CN)
QIAN HAOTIAN (CN)
JIANG TIANZI (CN)
SUN CHAOLIANG (CN)
WANG ZHICHAO (CN)
ZHANG HUAN (CN)
QIAN HAOTIAN (CN)
JIANG TIANZI (CN)
Application Number:
PCT/CN2023/124639
Publication Date:
April 25, 2024
Filing Date:
October 16, 2023
Export Citation:
Assignee:
ZHEJIANG LAB (CN)
International Classes:
G06T7/00; A61B5/00; A61B5/055; G06V10/82
Domestic Patent References:
WO2022147871A1 | 2022-07-14 |
Foreign References:
CN115359045A | 2022-11-18 | |||
CN115116607A | 2022-09-27 | |||
CN113539435A | 2021-10-22 | |||
CN107658018A | 2018-02-02 |
Attorney, Agent or Firm:
HANGZHOU QIUSHI PATENT OFFICE CO., LTD. (CN)
Download PDF:
Previous Patent: METHOD AND DEVICE FOR WIRELESS COMMUNICATIONS
Next Patent: SYSTEM FOR ACCURATE POSITIONING OF ABNORMAL REGIONS OF BRAIN FIBER TRACTS
Next Patent: SYSTEM FOR ACCURATE POSITIONING OF ABNORMAL REGIONS OF BRAIN FIBER TRACTS