Title
Modality alignment contrastive learning for severity assessment of COVID-19 from lung ultrasound and clinical information
Abstract
•We conducted the first large scale research for AI-based LUS analysis of COVID-19 and built an effective framework for COVID-19 diagnosis.•We designed a DSA-MIL module to combine the heterogeneous LUS images and videos.•We proposed a MA-CLR module to align the feature of LUS data and clinical information for patient severity prediction.•We learned the framework following an SRT pipeline to explore discriminative features gradually from different level.
Year
DOI
Venue
2021
10.1016/j.media.2021.101975
Medical Image Analysis
Keywords
DocType
Volume
Lung ultrasound,Multiple instance learning,Multi-modality,Contrastive learning
Journal
69
ISSN
Citations 
PageRank 
1361-8415
3
0.40
References 
Authors
0
25
Name
Order
Citations
PageRank
Wufeng Xue135018.80
Chunyan Cao230.40
Jie Liu330.40
Yilian Duan430.40
Haiyan Cao530.40
Jian Wang631.76
Xumin Tao730.40
Zejian Chen831.76
Meng Wu930.40
Jinxiang Zhang1030.40
Hui Sun1130.40
Yang Jin1230.40
Xin Yang13799.59
Ruobing Huang1433.78
Feixiang Xiang1530.40
Yue Song1630.40
Manjie You1730.40
Wen Zhang1851.74
Lili Jiang1930.40
Ziming Zhang2030.40
Shuangshuang Kong2130.40
Ying Tian2230.40
Li Zhang23183.04
Dong Ni2436737.37
Mingxing Xie2530.74