Batch Similarity Based Triplet Loss Assembled into Light-Weighted Convolutional Neural Networks for Medical Image Classification
In many medical image classification tasks, there is insufficient image data for deep convolutional neural networks (CNNs) to overcome the over-fitting problem. The light-weighted CNNs are easy to train but they usually have relatively poor classification performance. To improve the classification a...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-01-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/21/3/764 |
| _version_ | 1797407891511050240 |
|---|---|
| author | Zhiwen Huang Quan Zhou Xingxing Zhu Xuming Zhang |
| author_facet | Zhiwen Huang Quan Zhou Xingxing Zhu Xuming Zhang |
| author_sort | Zhiwen Huang |
| collection | DOAJ |
| description | In many medical image classification tasks, there is insufficient image data for deep convolutional neural networks (CNNs) to overcome the over-fitting problem. The light-weighted CNNs are easy to train but they usually have relatively poor classification performance. To improve the classification ability of light-weighted CNN models, we have proposed a novel batch similarity-based triplet loss to guide the CNNs to learn the weights. The proposed loss utilizes the similarity among multiple samples in the input batches to evaluate the distribution of training data. Reducing the proposed loss can increase the similarity among images of the same category and reduce the similarity among images of different categories. Besides this, it can be easily assembled into regular CNNs. To appreciate the performance of the proposed loss, some experiments have been done on chest X-ray images and skin rash images to compare it with several losses based on such popular light-weighted CNN models as EfficientNet, MobileNet, ShuffleNet and PeleeNet. The results demonstrate the applicability and effectiveness of our method in terms of classification accuracy, sensitivity and specificity. |
| first_indexed | 2024-03-09T03:50:21Z |
| format | Article |
| id | doaj.art-43b355cdf5504364bc2fd921efed80f4 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-09T03:50:21Z |
| publishDate | 2021-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-43b355cdf5504364bc2fd921efed80f42023-12-03T14:28:40ZengMDPI AGSensors1424-82202021-01-0121376410.3390/s21030764Batch Similarity Based Triplet Loss Assembled into Light-Weighted Convolutional Neural Networks for Medical Image ClassificationZhiwen Huang0Quan Zhou1Xingxing Zhu2Xuming Zhang3Key Laboratory of Molecular Biophysics of Ministry of Education, Department of Biomedical Engineering, School of Life Science and Technology, Huazhong University of Science and Technology, No 1037, Luoyu Road, Wuhan 430074, ChinaKey Laboratory of Molecular Biophysics of Ministry of Education, Department of Biomedical Engineering, School of Life Science and Technology, Huazhong University of Science and Technology, No 1037, Luoyu Road, Wuhan 430074, ChinaKey Laboratory of Molecular Biophysics of Ministry of Education, Department of Biomedical Engineering, School of Life Science and Technology, Huazhong University of Science and Technology, No 1037, Luoyu Road, Wuhan 430074, ChinaKey Laboratory of Molecular Biophysics of Ministry of Education, Department of Biomedical Engineering, School of Life Science and Technology, Huazhong University of Science and Technology, No 1037, Luoyu Road, Wuhan 430074, ChinaIn many medical image classification tasks, there is insufficient image data for deep convolutional neural networks (CNNs) to overcome the over-fitting problem. The light-weighted CNNs are easy to train but they usually have relatively poor classification performance. To improve the classification ability of light-weighted CNN models, we have proposed a novel batch similarity-based triplet loss to guide the CNNs to learn the weights. The proposed loss utilizes the similarity among multiple samples in the input batches to evaluate the distribution of training data. Reducing the proposed loss can increase the similarity among images of the same category and reduce the similarity among images of different categories. Besides this, it can be easily assembled into regular CNNs. To appreciate the performance of the proposed loss, some experiments have been done on chest X-ray images and skin rash images to compare it with several losses based on such popular light-weighted CNN models as EfficientNet, MobileNet, ShuffleNet and PeleeNet. The results demonstrate the applicability and effectiveness of our method in terms of classification accuracy, sensitivity and specificity.https://www.mdpi.com/1424-8220/21/3/764medical image classificationconvolutional neural networksbatch similarity based triplet loss |
| spellingShingle | Zhiwen Huang Quan Zhou Xingxing Zhu Xuming Zhang Batch Similarity Based Triplet Loss Assembled into Light-Weighted Convolutional Neural Networks for Medical Image Classification Sensors medical image classification convolutional neural networks batch similarity based triplet loss |
| title | Batch Similarity Based Triplet Loss Assembled into Light-Weighted Convolutional Neural Networks for Medical Image Classification |
| title_full | Batch Similarity Based Triplet Loss Assembled into Light-Weighted Convolutional Neural Networks for Medical Image Classification |
| title_fullStr | Batch Similarity Based Triplet Loss Assembled into Light-Weighted Convolutional Neural Networks for Medical Image Classification |
| title_full_unstemmed | Batch Similarity Based Triplet Loss Assembled into Light-Weighted Convolutional Neural Networks for Medical Image Classification |
| title_short | Batch Similarity Based Triplet Loss Assembled into Light-Weighted Convolutional Neural Networks for Medical Image Classification |
| title_sort | batch similarity based triplet loss assembled into light weighted convolutional neural networks for medical image classification |
| topic | medical image classification convolutional neural networks batch similarity based triplet loss |
| url | https://www.mdpi.com/1424-8220/21/3/764 |
| work_keys_str_mv | AT zhiwenhuang batchsimilaritybasedtripletlossassembledintolightweightedconvolutionalneuralnetworksformedicalimageclassification AT quanzhou batchsimilaritybasedtripletlossassembledintolightweightedconvolutionalneuralnetworksformedicalimageclassification AT xingxingzhu batchsimilaritybasedtripletlossassembledintolightweightedconvolutionalneuralnetworksformedicalimageclassification AT xumingzhang batchsimilaritybasedtripletlossassembledintolightweightedconvolutionalneuralnetworksformedicalimageclassification |