Title | ||
---|---|---|
Automatic Grading Assessments for Knee MRI Cartilage Defects via Self-ensembling Semi-supervised Learning with Dual-Consistency |
Abstract | ||
---|---|---|
•A semi-supervised learning framework called Dual-Consistency Mean Teacher (DC-MT) is designed for the knee cartilage defect assessment task.•A novel attention loss is defined to make the DC-MT model focus on the cartilage regions, yielding accurate attention masks for knee cartilage defect assessment.•A novel dual-consistency strategy is developed in the DC-MT model to penalize the inconsistent attention masks obtained from the student and the teacher sub-networks.•An aggregation network is proposed to ensemble features from all the slice-level by the DC-MT model, and finally yields the subject-level diagnosis. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1016/j.media.2022.102508 | Medical Image Analysis |
Keywords | DocType | Volume |
Knee cartilage defect,Semi-supervised learning,Dual consistency,Attention mechanism | Journal | 80 |
ISSN | Citations | PageRank |
1361-8415 | 0 | 0.34 |
References | Authors | |
3 | 13 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jiayu Huo | 1 | 24 | 4.91 |
Xi Ouyang | 2 | 0 | 1.01 |
Liping Si | 3 | 0 | 0.68 |
Kai Xuan | 4 | 0 | 1.35 |
Sheng Wang | 5 | 0 | 1.35 |
Weiwu Yao | 6 | 0 | 1.35 |
Ying Liu | 7 | 17 | 19.81 |
Jia Xu | 8 | 0 | 0.34 |
Dahong Qian | 9 | 27 | 7.78 |
Zhong Xue | 10 | 496 | 45.70 |
Qian Wang | 11 | 536 | 54.97 |
Dinggang Shen | 12 | 7837 | 611.27 |
Lichi Zhang | 13 | 23 | 7.12 |