赣南医科大学学报2025，Vol.45Issue(1)：68-73,6.DOI:10.3969/j.issn.1001-5779.2025.01.011
深度学习在膝关节MRI图像分割中的研究进展
Research progress of knee joint MRI image segmentation based on deep learning
向煜航 1孟滔 2刘佳奇 3张杏林4
作者信息
- 1. 赣南医科大学医学信息工程学院,江西 赣州 341000
- 2. 江西一脉阳光集团股份有限公司,江西 南昌 330000
- 3. 上海影禾医脉智能科技有限公司,上海 200336
- 4. 上海影禾医脉智能科技有限公司,上海 200336||滑铁卢大学大数据研究实验室,滑铁卢 N2L3G1
- 折叠
摘要
Abstract
With an increase in ageing population,the incidence of chronic joint diseases such as knee osteoarthritis(KOA)has increased year by year.Automatic segmentation of knee joint MRI images can significantly enhance the efficiency of diagnosis and treatment,reducing the burden on physicians.Traditional MRI image segmentation techniques are challenged by cumbersome processes,low precision,and limited generalization capabilities.In recent years,deep learning has shown great potential and advantages in medical image analysis.In knee joint MRI image segmentation,the DSC for bone can reach 0.98,and the DSC for all structures is more than 0.7.This paper reviews the research progress of deep learning in knee joint MRI image segmentation,focusing on the achievements and challenges of deep learning models in five different segmentation tasks.Although considerable progress has been achieved in the field of knee joint segmentation research,some challenges remain to be addressed for its full clinical application,including data diversity,model generalization,real-time performance,and standardization.关键词
膝关节/深度学习/图像分割/磁共振成像Key words
Knee joint/Deep learning/Image segmentation/Magnetic resonance imaging分类
临床医学引用本文复制引用
向煜航,孟滔,刘佳奇,张杏林..深度学习在膝关节MRI图像分割中的研究进展[J].赣南医科大学学报,2025,45(1):68-73,6.
