Title:
PHYSICS-BASED DATA AUGMENTATION CONTRASTIVE LEARNING REPRESENTATION IMAGING METHOD
Document Type and Number:
WIPO Patent Application WO/2023/115339
Kind Code:
A1
Abstract:
Disclosed in the present application are a physics-based data augmentation contrastive learning representation imaging method and apparatus, and a device and a storage medium thereof. The method comprises: performing undersampling data augmentation of multiple physical trajectories on magnetic resonance data; applying a contrastive representation learning framework to augmented undersampled magnetic resonance data; and ensuring the accuracy of magnetic resonance imaging by using a contrastive loss constraint and a data fitting item. By means of the solution provided in the present application, the problem of dependency on fully sampled data is solved, thereby improving the utilization efficiency of undersampled data. In addition, provided in the present invention is a parallel network framework for fast magnetic resonance imaging, thereby providing guidance for a magnetic resonance imaging technique.
More Like This:
Inventors:
WANG SHANSHAN (CN)
ZHENG HAIRONG (CN)
WU RUOYOU (CN)
LIU XIN (CN)
LIANG DONG (CN)
ZHENG HAIRONG (CN)
WU RUOYOU (CN)
LIU XIN (CN)
LIANG DONG (CN)
Application Number:
PCT/CN2021/140119
Publication Date:
June 29, 2023
Filing Date:
December 21, 2021
Export Citation:
Assignee:
SHENZHEN INST ADV TECH (CN)
International Classes:
G06T11/00; A61B5/055; G06N3/08
Foreign References:
| CN113192151A | 2021-07-30 | |||
| CN110992440A | 2020-04-10 | |||
| CN106491131A | 2017-03-15 |
Other References:
CHENG HUI-TAO, WANG SHAN-SHAN, KE ZI-WEN, JIA SEN, CHENG JING, QIU ZHI-LANG, ZHENG HAI-RONG, LIANG DONG: "A Deep Recursive Cascaded Convolutional Network for Parallel MRI", CHINESE JOURNAL OF MAGNETIC RESONANCE, vol. 36, no. 4, 15 December 2019 (2019-12-15), CN , pages 437 - 445, XP093068441, ISSN: 1000-4556, DOI: 10.11938/cjmr20192721
Attorney, Agent or Firm:
BEIJING CHENGHUI LAW FIRM (CN)
Download PDF: