Name
Title | ||
|---|---|---|
Contrastive Learning Based Intelligent Skin Lesion Diagnosis in Edge Computing Networks |
Abstract | ||
|---|---|---|
In recent years, automatic skin lesion diagnosis methods based on artificial intelligence (AI) have achieved great success. However, the lack of rich available training data and visual similarity of various skin diseases remain the major challenges in intelligent skin lesion diagnosis. In this paper, we propose a contrastive learning based intelligent skin lesion diagnosis (CL-ISLD) scheme in edge computing networks. Specifically, an edge computing based intelligent skin lesion diagnosis network is constructed, which can provide the convenient and quick online diagnosis service to users nearby. Meanwhile, a contrastive learning based dual encoder network is designed to overcome training sample scarcity by fully leveraging unlabeled samples for performance promotion. Subsequently, we devise a maximum mean discrepancy (MMD) based supervised contrastive loss function, it can efficiently explore complex intra-class and inter-class variances of samples. Finally, the simulation results demonstrate that the proposed CL-ISLD obtains competitive diagnosis accuracy compared with existing representative works and achieves relatively more balanced performance among classes in inadequate and imbalanced dataset. |
Year | DOI | Venue |
|---|---|---|
2021 | 10.1109/GLOBECOM46510.2021.9685609 | 2021 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM) |
Keywords | DocType | ISSN |
Skin lesion, contrastive learning, edge computing, loss function | Conference | 2334-0983 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Yanhang Shi | 1 | 0 | 0.34 |
| Congying Duan | 2 | 0 | 0.34 |
| Siguang Chen | 3 | 63 | 12.91 |