Abstract | ||
---|---|---|
•A novel dual-curriculum learning paradigm (SEDC) is proposed for the first time to handle two types of data imbalances in glaucoma diagnosis by feature space augmenting.•An effective self-ensemble learning framework is developed to reinforce the discriminative ability of feature representation for the rare cases by feature distillation.•A contrastive re-balanced loss is constructed to jointly learn the discriminative representation and the powerful classifier by integrating supervised contrastive learning into the sample re-balancing strategy. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1016/j.media.2021.102295 | Medical Image Analysis |
Keywords | DocType | Volume |
Curriculum learning,Glaucoma diagnosis,Data imbalance,Feature augmentation,Computer-aided diagnosis,Self ensembling | Journal | 75 |
ISSN | Citations | PageRank |
1361-8415 | 0 | 0.34 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Rongchang Zhao | 1 | 30 | 4.63 |
Xuanlin Chen | 2 | 0 | 0.34 |
Zailiang Chen | 3 | 43 | 9.10 |
Shuo Li | 4 | 887 | 72.47 |