journal article·2024-11-15Open access

A New Image Oversampling Method Based on Influence Functions and Weights

Jun Ye, Shoulei Lu, Jiawei Chen

Applied Sciences· Volume 14· Issue 22

Abstract

Although imbalanced data have been studied for many years, the problem of data imbalance is still a major problem in the development of machine learning and artificial intelligence. The development of deep learning and artificial intelligence has further expanded the impact of imbalanced data, so studying imbalanced data classification is of practical significance. We propose an image oversampling algorithm based on the influence function and sample weights. Our scheme not only synthesizes high-quality minority class samples but also preserves the original features and information of minority class images. To address the lack of visually reasonable features in SMOTE when synthesizing images, we improve the pre-training model by removing the pooling layer and the fully connected layer in the model, extracting the important features of the image by convolving the image, executing SMOTE interpolation operation on the extracted important features to derive the synthesized image features, and inputting the features into a DCGAN network generator, which maps these features into the high-dimensional image space to generate a realistic image. To verify that our scheme can synthesize high-quality images and thus improve classification accuracy, we conduct experiments on the processed CIFAR10, CIFAR100, and ImageNet-LT datasets.
Source: crossref · Quality: baseline

Research context

OpenAlex topics: Image Processing Techniques and Applications · Image and Object Detection Techniques · Advanced Measurement and Detection Methods

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

Key Laboratory of Internet Information Retrieval of Hainan Province, School of Cyberspace Security, Hainan University, Haikou 570228, China

School of Cyberspace Security, Hainan University, Haikou 570228, China

Funding

National Natural Science Foundation of China

References (37)

Show source-provided references
1. Learning from class-imbalanced data: Review of data driven methods and algorithm driven methods10.3934/DSFE.2021002 ↗
2. Tarekegn, A.N., Giacobini, M., and Michalak, K. (2021). A review of methods for imbalanced multi-label classification. Pattern Recognit., 118.10.1016/j.patcog.2021.107965 ↗
3. Wu, Z., Guo, K., Luo, E., Wang, T., Wang, S., Yang, Y., Zhu, X., and Ding, R. (2024). Medical long-tailed learning for imbalanced data: Bibliometric analysis. Comput. Methods Programs Biomed.10.1016/j.cmpb.2024.108106 ↗
4. Yuan, X., Chen, S., Sun, C., and Yuwen, L. (2021, January 9–12). A novel class imbalance-oriented polynomial neural network algorithm for disease diagnosis. Proceedings of the 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), Houston, TX, USA.10.1109/BIBM52615.2021.9669281 ↗
5. A survey on GAN techniques for data augmentation to address the imbalanced data issues in credit card fraud detection10.3390/make5010019 ↗
6. Balla, A., Habaebi, M.H., Elsheikh, E.A.A., Islam, M.R., and Suliman, F.M. (2023). The effect of dataset imbalance on the performance of SCADA intrusion detection systems. Sensors, 23.10.3390/s23020758 ↗
7. A theoretical distribution analysis of synthetic minority oversampling technique (SMOTE) for imbalanced learning
8. Wongvorachan, T., He, S., and Bulut, O. (2023). A comparison of undersampling, oversampling, and SMOTE methods for dealing with imbalanced classification in educational data mining. Information, 14.10.3390/info14010054 ↗
9. Qing, Z., Zeng, Q., Wang, H., Liu, Y., Xiong, T., and Zhang, S. (2022). ADASYN-LOF Algorithm for imbalanced tornado samples. Atmosphere, 13.10.3390/atmos13040544 ↗
10. RCSMOTE: Range-controlled synthetic minority over-sampling technique for handling the class imbalance problem10.1016/j.ins.2020.07.014 ↗
11. Dixit, A., and Mani, A. (2023). Sampling technique for noisy and borderline examples problem in imbalanced classification. Appl. Soft Comput., 142.10.1016/j.asoc.2023.110361 ↗
12. DeepSMOTE: Fusing deep learning and SMOTE for imbalanced data
13. Maldonado, S., Vairetti, C., Fernandez, A., and Herrera, F. (2022). FW-SMOTE: A feature-weighted oversampling approach for imbalanced classification. Pattern Recognit., 124.10.1016/j.patcog.2021.108511 ↗
14. Nguyen, T., Mengersen, K., Sous, D., and Liquet, B. (2023). SMOTE-CD: SMOTE for compositional data. PLoS ONE, 18.10.1371/journal.pone.0287705 ↗
15. Generative adversarial nets
16. Effective data generation for imbalanced learning using conditional generative adversarial networks10.1016/j.eswa.2017.09.030 ↗
17. Mariani, G., Scheidegger, F., Istrate, R., Bekas, C., and Malossi, C. (2018). Bagan: Data augmentation with balancing GAN. arXiv.
18. SMOTified-GAN for class imbalanced pattern classification problems10.1109/ACCESS.2022.3158977 ↗
19. Park, S., Lim, J., Jeon, Y., and Choi, J.Y. (2021, January 10–17). Influence-balanced loss for imbalanced visual classification. Proceedings of the IEEE/CVF International Conference on Computer Vision, Montreal, QC, Canada.10.1109/ICCV48922.2021.00077 ↗
20. Yeung, M., Sala, E., Schönlieb, C.-B., and Rundo, L. (2022). Unified focal loss: Generalising dice and cross entropy-based losses to handle class imbalanced medical image segmentation. Comput. Med Imaging Graph., 95.10.1016/j.compmedimag.2021.102026 ↗
21. Neural collapse inspired attraction–repulsion-balanced loss for imbalanced learning10.1016/j.neucom.2023.01.023 ↗
22. PCCT: Progressive class-center triplet loss for imbalanced medical image classification10.1109/JBHI.2023.3240136 ↗
23. Cost-sensitive learning with modified Stein loss function10.1016/j.neucom.2023.01.052 ↗
24. Cui, Y., Jia, M., Lin, T.-Y., Song, Y., and Belongie, S. (2019, January 16–20). Class-balanced loss based on effective number of samples. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, CA, USA.10.1109/CVPR.2019.00949 ↗
25. Learning imbalanced datasets with label-distribution-aware margin loss
26. Gidaris, S., and Komodakis, N. (2018, January 18–22). Dynamic few-shot visual learning without forgetting. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.10.1109/CVPR.2018.00459 ↗
27. Liu, Z., Miao, Z., Zhan, X., Wang, J., Gong, B., and Yu, S.X. (2019, January 15–20). Large-scale long-tailed recognition in an open world. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, CA, USA.10.1109/CVPR.2019.00264 ↗
28. Li, B., Yao, Y., Tan, J., Zhang, G., Yu, F., Lu, J., and Luo, Y. (2022, January 19–24). Equalized focal loss for dense long-tailed object detection. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, USA.10.1109/CVPR52688.2022.00686 ↗
29. Wei, H., Tao, L., Xie, R., Feng, L., and An, B. (2022, January 18–24). Open-sampling: Exploring out-of-distribution data for re-balancing long-tailed datasets. Proceedings of the International Conference on Machine Learning, Virtual.
30. He, K., Zhang, X., Ren, S., and Sun, J. (July, January 26). Deep residual learning for image recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA.
31. Combined cleaning and resampling algorithm for multi-class imbalanced data with label noise10.1016/j.knosys.2020.106223 ↗
32. Radial-based undersampling for imbalanced data classification10.1016/j.patcog.2020.107262 ↗
33. Xiang, L., Ding, G., and Han, J. (2021, January 5–6). Increasing oversampling diversity for long-tailed visual recognition. Proceedings of the Artificial Intelligence: First CAAI International Conference, CICAI 2021, Hangzhou, China. Part I.10.1007/978-3-030-93046-2_4 ↗
34. Guan, S., Zhao, X., Xue, Y., and Pan, H. (2024). AWGAN: An adaptive weighting GAN approach for oversampling imbalanced datasets. Inf. Sci., 663.10.1016/j.ins.2024.120311 ↗
35. Heroza, R.I., Gan, J.Q., and Raza, H. (2023, January 19–21). SIA-SMOTE: A SMOTE-based oversampling method with better interpolation on high-dimensional data by using a Siamese network. Proceedings of the International Work-Conference on Artificial Neural Networks, Ponta Delgada, Portugal.10.1007/978-3-031-43085-5_35 ↗
36. A GAN-based data augmentation method for imbalanced multi-class skin lesion classification10.1109/ACCESS.2024.3360215 ↗
37. Few-shot GAN: Improving the performance of intelligent fault diagnosis in severe data imbalance