Adversarial Autoencoders Oversampling Algorithm for Imbalanced Image Data

ZHI Weimei; CHANG Zhi; LU Junhua; GENG Zhengqian

doi:10.11999/JEIT240330

Article Contents

Article Navigation > Journal of Electronics & Information Technology > 2025 >

ZHI Weimei, CHANG Zhi, LU Junhua, GENG Zhengqian. Adversarial Autoencoders Oversampling Algorithm for Imbalanced Image Data[J]. Journal of Electronics & Information Technology. doi: 10.11999/JEIT240330

Citation:

ZHI Weimei, CHANG Zhi, LU Junhua, GENG Zhengqian. Adversarial Autoencoders Oversampling Algorithm for Imbalanced Image Data[J]. Journal of Electronics & Information Technology. doi: 10.11999/JEIT240330

Citation:

PDF( 6910 KB)

Adversarial Autoencoders Oversampling Algorithm for Imbalanced Image Data

doi: 10.11999/JEIT240330

School of Computer and Artificial Intelligence, Zhengzhou University, Zhengzhou 450001, China

Funds: The National Key Research and Development Project (2023YFC2206404)

Received Date: 2024-04-24
Rev Recd Date: 2024-09-18

Available Online: 2024-09-24

Abstract

Abstract

Many traditional imbalanced learning algorithms suitable for low-dimensional data do not perform well on image data. Although the oversampling algorithm based on Generative Adversarial Networks (GAN) can generate high-quality images, it is prone to mode collapse in the case of class imbalance. Oversampling algorithms based on AutoEncoders (AE) are easy to train, but the generated images are of lower quality. In order to improve the quality of samples generated by the oversampling algorithm in imbalanced images and the stability of training, a Balanced oversampling method with AutoEncoders and Generative Adversarial Networks (BAEGAN) is proposed by this paper, which is based on the idea of GAN and AE. First, a conditional embedding layer is introduced in the Autoencoder, and the pre-trained conditional Autoencoder is used to initialize the GAN to stabilize the model training; then the output structure of the discriminator is improved, and a loss function that combines Focal Loss and gradient penalty is proposed to alleviate the impact of class imbalance; and finally the Synthetic Minority Oversampling TEchnique (SMOTE) is used to generate high-quality images from the distribution map of latent vectors. Experimental results on four image data sets show that the proposed algorithm is superior to oversampling methods such as Auxiliary Classifier Generative Adversarial Networks (ACGAN) and BAlancing Generative Adversarial Networks (BAGAN) in terms of image quality and classification performance after oversampling and can effectively solve the class imbalance problem in image data.

FullText(HTML)

References(26)

References

[1]	FAN Xi, GUO Xin, CHEN Qi, et al. Data augmentation of credit default swap transactions based on a sequence GAN[J]. Information Processing & Management, 2022, 59(3): 102889. doi: 10.1016/j.ipm.2022.102889.
[2]	刘侠, 吕志伟, 李博, 等. 基于多尺度残差双域注意力网络的乳腺动态对比度增强磁共振成像肿瘤分割方法[J]. 电子与信息学报, 2023, 45(5): 1774–1785. doi: 10.11999/JEIT220362. LIU Xia, LÜ Zhiwei, LI Bo, et al. Segmentation algorithm of breast tumor in dynamic contrast-enhanced magnetic resonance imaging based on network with multi-scale residuals and dual-domain attention[J]. Journal of Electronics & Information Technology, 2023, 45(5): 1774–1785. doi: 10.11999/JEIT220362.
[3]	尹梓诺, 马海龙, 胡涛. 基于联合注意力机制和一维卷积神经网络-双向长短期记忆网络模型的流量异常检测方法[J]. 电子与信息学报, 2023, 45(10): 3719–3728. doi: 10.11999/JEIT220959. YIN Zinuo, MA Hailong, and HU Tao. A traffic anomaly detection method based on the joint model of attention mechanism and one-dimensional convolutional neural network-bidirectional long short term memory[J]. Journal of Electronics & Information Technology, 2023, 45(10): 3719–3728. doi: 10.11999/JEIT220959.
[4]	FERNÁNDEZ A, GARCÍA S, GALAR M, et al. Learning From Imbalanced Data Sets[M]. Cham: Springer, 2018: 327–349. doi: 10.1007/978-3-319-98074-4.
[5]	HUANG Zhan’ao, SANG Yongsheng, SUN Yanan, et al. A neural network learning algorithm for highly imbalanced data classification[J]. Information Sciences, 2022, 612: 496–513. doi: 10.1016/j.ins.2022.08.074.
[6]	FU Saiji, YU Xiaotong, and TIAN Yingjie. Cost sensitive ν-support vector machine with LINEX loss[J]. Information Processing & Management, 2022, 59(2): 102809. doi: 10.1016/j.ipm.2021.102809.
[7]	LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]. The IEEE International Conference on Computer Vision, Venice, Italy, 2017: 2999–3007. doi: 10.1109/ICCV.2017.324.
[8]	LI Buyu, LIU Yu, and WANG Xiaogang. Gradient harmonized single-stage detector[C]. The 33rd AAAI Conference on Artificial Intelligence, Honolulu, USA, 2019: 8577–8584. doi: 10.1609/aaai.v33i01.33018577.
[9]	MICHELUCCI U. An introduction to autoencoders[J]. arXiv preprint arXiv: 2201.03898, 2022. doi: 10.48550/arXiv.2201.03898.
[10]	GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial nets[C]. The 27th International Conference on Neural Information Processing Systems, Montreal, Canada, 2014: 2672–2680.
[11]	DABLAIN D, KRAWCZYK B, and CHAWLA N V. DeepSMOTE: Fusing deep learning and SMOTE for imbalanced data[J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(9): 6390–6404. doi: 10.1109/TNNLS.2021.3136503.
[12]	MIRZA M and OSINDERO S. Conditional generative adversarial nets[J]. arXiv preprint arXiv: 1411.1784, 2014. doi: 10.48550/arXiv.1411.1784.
[13]	ODENA A, OLAH C, and SHLENS J. Conditional image synthesis with auxiliary classifier GANs[C]. The 34th International Conference on Machine Learning, Sydney, Australia, 2017: 2642–2651.
[14]	GULRAJANI I, AHMED F, ARJOVSKY M, et al. Improved training of wasserstein GANs[C]. The 31st International Conference on Neural Information Processing Systems, Long Beach, USA, 2017: 5769–5779.
[15]	CHAWLA N V, BOWYER K W, HALL L O, et al. SMOTE: Synthetic minority over-sampling technique[J]. Journal of Artificial Intelligence Research, 2002, 16: 321–357. doi: 10.1613/jair.953.
[16]	MARIANI G, SCHEIDEGGER F, ISTRATE R, et al. BAGAN: Data augmentation with balancing GAN[J]. arXiv preprint arXiv: 1803.09655, 2018. doi: 10.48550/arXiv.1803.09655.
[17]	HUANG Gaofeng and JAFARI A H. Enhanced balancing GAN: Minority-class image generation[J]. Neural Computing and Applications, 2023, 35(7): 5145–5154. doi: 10.1007/s00521-021-06163-8.
[18]	BAO Jianmin, CHEN Dong, WEN Fang, et al. CVAE-GAN: Fine-grained image generation through asymmetric training[C]. The IEEE International Conference on Computer Vision (ICCV), Venice, Italy, 2017: 2764–2773. doi: 10.1109/ICCV.2017.299.
[19]	MAKHZANI A, SHLENS J, JAITLY N, et al. Adversarial autoencoders[J]. arXiv preprint arXiv: 1511.05644, 2015. doi: 10.48550/arXiv.1511.05644.
[20]	CUI Yin, JIA Menglin, LIN T Y, et al. Class-balanced loss based on effective number of samples[C]. The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, USA, 2019: 9260–9269. doi: 10.1109/CVPR.2019.00949.
[21]	DOWSON D C and LANDAU B V. The Fréchet distance between multivariate normal distributions[J]. Journal of Multivariate Analysis, 1982, 12(3): 450–455. doi: 10.1016/0047-259X(82)90077-X.
[22]	HUANG Chen, LI Yining, LOY C C, et al. Learning deep representation for imbalanced classification[C]. The IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, USA, 2016: 5375–5384. doi: 10.1109/CVPR.2016.580.
[23]	KUBAT M and MATWIN S. Addressing the curse of imbalanced training sets: One-sided selection[C]. 14th International Conference on Machine Learning, Nashville, USA, 1997: 179–186.
[24]	HRIPCSAK G and ROTHSCHILD A S. Agreement, the F-measure, and reliability in information retrieval[J]. Journal of the American Medical Informatics Association, 2005, 12(3): 296–298. doi: 10.1197/jamia.M1733.
[25]	SOKOLOVA M and LAPALME G. A systematic analysis of performance measures for classification tasks[J]. Information Processing & Management, 2009, 45(4): 427–437. doi: 10.1016/j.ipm.2009.03.002.
[26]	RADFORD A, METZ L, and CHINTALA S. Unsupervised representation learning with deep convolutional generative adversarial networks[C]. 4th International Conference on Learning Representations, San Juan, Puerto Rico, 2016.