Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant Loss
Melanoma recognition is challenging due to data imbalance and high intra-class variations and large inter-class similarity. Aiming at the issues, we propose a melanoma recognition method using deep convolutional neural network with covariance discriminant loss in dermoscopy images. Deep convolutiona...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-10-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/20/20/5786 |
| _version_ | 1797551119931539456 |
|---|---|
| author | Lei Guo Gang Xie Xinying Xu Jinchang Ren |
| author_facet | Lei Guo Gang Xie Xinying Xu Jinchang Ren |
| author_sort | Lei Guo |
| collection | DOAJ |
| description | Melanoma recognition is challenging due to data imbalance and high intra-class variations and large inter-class similarity. Aiming at the issues, we propose a melanoma recognition method using deep convolutional neural network with covariance discriminant loss in dermoscopy images. Deep convolutional neural network is trained under the joint supervision of cross entropy loss and covariance discriminant loss, rectifying the model outputs and the extracted features simultaneously. Specifically, we design an embedding loss, namely covariance discriminant loss, which takes the first and second distance into account simultaneously for providing more constraints. By constraining the distance between hard samples and minority class center, the deep features of melanoma and non-melanoma can be separated effectively. To mine the hard samples, we also design the corresponding algorithm. Further, we analyze the relationship between the proposed loss and other losses. On the International Symposium on Biomedical Imaging (ISBI) 2018 Skin Lesion Analysis dataset, the two schemes in the proposed method can yield a sensitivity of 0.942 and 0.917, respectively. The comprehensive results have demonstrated the efficacy of the designed embedding loss and the proposed methodology. |
| first_indexed | 2024-03-10T15:40:10Z |
| format | Article |
| id | doaj.art-2330db6b78604af4a7c6abc7ca4c980e |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T15:40:10Z |
| publishDate | 2020-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-2330db6b78604af4a7c6abc7ca4c980e2023-11-20T16:52:21ZengMDPI AGSensors1424-82202020-10-012020578610.3390/s20205786Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant LossLei Guo0Gang Xie1Xinying Xu2Jinchang Ren3College of Information and Computer, Taiyuan University of Technology, Taiyuan 030024, ChinaCollege of Electrical and Power Engineering, Taiyuan University of Technology, Taiyuan 030024, ChinaCollege of Electrical and Power Engineering, Taiyuan University of Technology, Taiyuan 030024, ChinaCollege of Electrical and Power Engineering, Taiyuan University of Technology, Taiyuan 030024, ChinaMelanoma recognition is challenging due to data imbalance and high intra-class variations and large inter-class similarity. Aiming at the issues, we propose a melanoma recognition method using deep convolutional neural network with covariance discriminant loss in dermoscopy images. Deep convolutional neural network is trained under the joint supervision of cross entropy loss and covariance discriminant loss, rectifying the model outputs and the extracted features simultaneously. Specifically, we design an embedding loss, namely covariance discriminant loss, which takes the first and second distance into account simultaneously for providing more constraints. By constraining the distance between hard samples and minority class center, the deep features of melanoma and non-melanoma can be separated effectively. To mine the hard samples, we also design the corresponding algorithm. Further, we analyze the relationship between the proposed loss and other losses. On the International Symposium on Biomedical Imaging (ISBI) 2018 Skin Lesion Analysis dataset, the two schemes in the proposed method can yield a sensitivity of 0.942 and 0.917, respectively. The comprehensive results have demonstrated the efficacy of the designed embedding loss and the proposed methodology.https://www.mdpi.com/1424-8220/20/20/5786melanoma recognitionembedding losscovariance discriminant lossdeep convolutional neural networkdermoscopy image |
| spellingShingle | Lei Guo Gang Xie Xinying Xu Jinchang Ren Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant Loss Sensors melanoma recognition embedding loss covariance discriminant loss deep convolutional neural network dermoscopy image |
| title | Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant Loss |
| title_full | Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant Loss |
| title_fullStr | Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant Loss |
| title_full_unstemmed | Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant Loss |
| title_short | Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant Loss |
| title_sort | effective melanoma recognition using deep convolutional neural network with covariance discriminant loss |
| topic | melanoma recognition embedding loss covariance discriminant loss deep convolutional neural network dermoscopy image |
| url | https://www.mdpi.com/1424-8220/20/20/5786 |
| work_keys_str_mv | AT leiguo effectivemelanomarecognitionusingdeepconvolutionalneuralnetworkwithcovariancediscriminantloss AT gangxie effectivemelanomarecognitionusingdeepconvolutionalneuralnetworkwithcovariancediscriminantloss AT xinyingxu effectivemelanomarecognitionusingdeepconvolutionalneuralnetworkwithcovariancediscriminantloss AT jinchangren effectivemelanomarecognitionusingdeepconvolutionalneuralnetworkwithcovariancediscriminantloss |