{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,6]],"date-time":"2026-02-06T18:16:53Z","timestamp":1770401813937,"version":"3.49.0"},"reference-count":44,"publisher":"PeerJ","license":[{"start":{"date-parts":[[2026,2,6]],"date-time":"2026-02-06T00:00:00Z","timestamp":1770336000000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Development and Promotion of Science and Technology Talents Project (DPST), Royal Government of Thailand Scholarship","award":["592002"],"award-info":[{"award-number":["592002"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>Deep learning models often exhibit bias when trained on imbalanced datasets, tending to favor majority classes. While Deep Synthetic Minority Over-sampling Technique (DeepSMOTE) mitigates this challenge by generating synthetic minority class images within an autoencoder\u2019s latent space, the proposed architecture achieves further performance enhancement through the optimization of the latent space structure. This strategy introduces a composite loss function that minimizes intra-class distances while maximizing inter-class distances by incorporating class centroid information into the autoencoder training process. The approach was evaluated on six benchmark image datasets (MNIST, Fashion-MNIST, EMNIST, SVHN, GTSRB, and CIFAR10) with severe class imbalance ratios of 750:1, using five-fold cross-validation. Results demonstrate that Deep Class-Latent Synthetic Minority Oversampling Technique (DeepCLSMOTE) consistently outperforms baseline methods, including Balancing Generative Adversarial Network (BAGAN) and DeepSMOTE across all evaluation metrics, achieving statistically significant improvements in macro-average accuracy, precision, recall, and F1-measure. The enhanced performance is attributed to improved discriminative feature extraction in the optimized latent space, resulting in superior classification performance on highly imbalanced image datasets, particularly for critical minority classes.<\/jats:p>","DOI":"10.7717\/peerj-cs.3461","type":"journal-article","created":{"date-parts":[[2026,2,6]],"date-time":"2026-02-06T08:03:51Z","timestamp":1770365031000},"page":"e3461","source":"Crossref","is-referenced-by-count":0,"title":["DeepCLSMOTE: deep class-latent synthetic minority oversampling technique"],"prefix":"10.7717","volume":"12","author":[{"given":"Suvaporn","family":"Homjandee","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9733-8937","authenticated-orcid":true,"given":"Krung","family":"Sinapiromsaran","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"4443","published-online":{"date-parts":[[2026,2,6]]},"reference":[{"key":"10.7717\/peerj-cs.3461\/ref-1","first-page":"214","article-title":"Wasserstein generative adversarial networks","author":"Arjovsky","year":"2017"},{"issue":"3","key":"10.7717\/peerj-cs.3461\/ref-2","doi-asserted-by":"publisher","DOI":"10.1177\/1088467x241305509","article-title":"A review of intelligent data analysis: machine learning approaches for addressing class imbalance in healthcare\u2014challenges and perspectives","volume":"29","author":"Aubaidan","year":"2025","journal-title":"Intelligent Data Analysis"},{"issue":"7","key":"10.7717\/peerj-cs.3461\/ref-3","doi-asserted-by":"publisher","first-page":"249","DOI":"10.1016\/j.neunet.2018.07.011","article-title":"A systematic study of the class imbalance problem in convolutional neural networks","volume":"106","author":"Buda","year":"2018","journal-title":"Neural Networks"},{"key":"10.7717\/peerj-cs.3461\/ref-4","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1613\/jair.953","article-title":"Smote: synthetic minority over-sampling technique","volume":"16","author":"Chawla","year":"2002","journal-title":"Journal of Artificial Intelligence Research"},{"key":"10.7717\/peerj-cs.3461\/ref-5","first-page":"2921","article-title":"EMNIST: an extension of MNIST to handwritten letters","author":"Cohen","year":"2017"},{"issue":"9","key":"10.7717\/peerj-cs.3461\/ref-6","doi-asserted-by":"publisher","first-page":"6390","DOI":"10.1109\/tnnls.2021.3136503","article-title":"Deepsmote: fusing deep learning and smote for imbalanced data","volume":"34","author":"Dablain","year":"2023","journal-title":"IEEE Transactions on Neural Networks and Learning Systems"},{"key":"10.7717\/peerj-cs.3461\/ref-7","first-page":"1","article-title":"Statistical comparisons of classifiers over multiple data sets","volume":"7","author":"Dem\u0161ar","year":"2006","journal-title":"Journal of Machine Learning Research"},{"key":"10.7717\/peerj-cs.3461\/ref-8","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-319-98074-4","volume-title":"Learning from imbalanced data sets","author":"Fern\u00e1ndez","year":"2018"},{"key":"10.7717\/peerj-cs.3461\/ref-9","volume-title":"Deep learning","author":"Goodfellow","year":"2016"},{"key":"10.7717\/peerj-cs.3461\/ref-10","article-title":"Generative adversarial nets","volume":"27","author":"Goodfellow","year":"2014"},{"key":"10.7717\/peerj-cs.3461\/ref-11","article-title":"Metrics for multi-class classification: an overview","author":"Grandini","year":"2020"},{"key":"10.7717\/peerj-cs.3461\/ref-12","article-title":"Improved training of wasserstein gans","author":"Gulrajani","year":"2017"},{"issue":"3","key":"10.7717\/peerj-cs.3461\/ref-13","doi-asserted-by":"publisher","first-page":"2429","DOI":"10.1007\/s11063-022-10977-5","article-title":"A survey on adversarial domain adaptation","volume":"55","author":"Hassan Pour Zonoozi","year":"2022","journal-title":"Neural Processing Letters"},{"key":"10.7717\/peerj-cs.3461\/ref-14","first-page":"770","article-title":"Deep residual learning for image recognition","author":"He","year":"2016"},{"issue":"1","key":"10.7717\/peerj-cs.3461\/ref-15","doi-asserted-by":"publisher","first-page":"25","DOI":"10.18034\/ajase.v8i1.9","article-title":"Gans trained by a two time-scale update rule converge to a nash equilibrium","volume":"30","author":"Heusel","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"10.7717\/peerj-cs.3461\/ref-16","article-title":"DeepCLSMOTE: deep class-latent synthetic minority oversampling technique","author":"Homjandee","year":"2025"},{"key":"10.7717\/peerj-cs.3461\/ref-17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1502.03167","article-title":"Batch normalization: accelerating deep network training by reducing internal covariate shift","author":"Ioffe","year":"2015"},{"issue":"1","key":"10.7717\/peerj-cs.3461\/ref-18","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1186\/s40537-019-0192-5","article-title":"Survey on deep learning with class imbalance","volume":"6","author":"Johnson","year":"2019","journal-title":"Journal of Big Data"},{"issue":"6","key":"10.7717\/peerj-cs.3461\/ref-19","doi-asserted-by":"publisher","first-page":"4006","DOI":"10.3390\/app13064006","article-title":"Effective class-imbalance learning based on smote and convolutional neural networks","volume":"13","author":"Joloudari","year":"2023","journal-title":"Applied Sciences"},{"key":"10.7717\/peerj-cs.3461\/ref-20","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2304.02858","article-title":"A review of ensemble learning and data augmentation models for class imbalanced problems: combination, implementation and evaluation","author":"Khan","year":"2023"},{"key":"10.7717\/peerj-cs.3461\/ref-21","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1312.6114","article-title":"Auto-encoding variational bayes","author":"Kingma","year":"2013"},{"issue":"4","key":"10.7717\/peerj-cs.3461\/ref-22","doi-asserted-by":"publisher","first-page":"221","DOI":"10.1007\/s13748-016-0094-0","article-title":"Learning from imbalanced data: open challenges and future directions","volume":"5","author":"Krawczyk","year":"2016","journal-title":"Progress in Artificial Intelligence"},{"key":"10.7717\/peerj-cs.3461\/ref-23","article-title":"Learning multiple layers of features from tiny images","author":"Krizhevsky","year":"2009"},{"key":"10.7717\/peerj-cs.3461\/ref-24","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1601.04756","article-title":"Improved sampling techniques for learning an imbalanced data set","author":"Lauron","year":"2016"},{"key":"10.7717\/peerj-cs.3461\/ref-25","article-title":"The mnist database of handwritten digits","author":"LeCun","year":"1998"},{"key":"10.7717\/peerj-cs.3461\/ref-26","first-page":"2999","article-title":"Focal loss for dense object detection","author":"Lin","year":"2017"},{"key":"10.7717\/peerj-cs.3461\/ref-27","article-title":"BAGAN: data augmentation with balancing GAN","author":"Mariani","year":"2018"},{"key":"10.7717\/peerj-cs.3461\/ref-28","article-title":"Spectral normalization for generative adversarial networks","author":"Miyato","year":"2018"},{"key":"10.7717\/peerj-cs.3461\/ref-29","first-page":"337","volume-title":"Mean squared error","author":"MSE","year":"2008"},{"key":"10.7717\/peerj-cs.3461\/ref-30","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV.2019.00178","article-title":"Generative adversarial minority oversampling","author":"Mullick","year":"2019"},{"key":"10.7717\/peerj-cs.3461\/ref-31","article-title":"Reading digits in natural images with unsupervised feature learning","author":"Netzer","year":"2011"},{"key":"10.7717\/peerj-cs.3461\/ref-32","article-title":"Pytorch: an imperative style, high-performance deep learning library","author":"Paszke","year":"2019"},{"key":"10.7717\/peerj-cs.3461\/ref-33","article-title":"Improved techniques for training GANs","author":"Salimans","year":"2016"},{"issue":"2s","key":"10.7717\/peerj-cs.3461\/ref-34","doi-asserted-by":"publisher","first-page":"910","DOI":"10.52783\/jes.1688","article-title":"A comprehensive analysis of exploring the efficacy of machine learning algorithms in text, image, and speech analysis","volume":"20","author":"Sophia","year":"2024","journal-title":"Journal of Electrical Systems"},{"key":"10.7717\/peerj-cs.3461\/ref-35","first-page":"32","article-title":"Man vs. computer: benchmarking machine learning algorithms for traffic sign recognition","author":"Stallkamp","year":"2012"},{"key":"10.7717\/peerj-cs.3461\/ref-36","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR42600.2020.01168","article-title":"Equalization loss for long-tailed object recognition","author":"Tan","year":"2020"},{"issue":"C","key":"10.7717\/peerj-cs.3461\/ref-37","doi-asserted-by":"publisher","first-page":"112143","DOI":"10.1016\/j.knosys.2024.112143","article-title":"Robust two-stage instance-level cost-sensitive learning method for class imbalance problem","volume":"300","author":"Tang","year":"2024","journal-title":"Knowledge-Based Systems"},{"key":"10.7717\/peerj-cs.3461\/ref-38","article-title":"Loss functions and metrics in deep learning. A review","author":"Terven","year":"2023"},{"key":"10.7717\/peerj-cs.3461\/ref-39","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1601.06759","article-title":"Pixel recurrent neural networks","author":"van den Oord","year":"2016"},{"key":"10.7717\/peerj-cs.3461\/ref-40","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1708.07747","article-title":"Fashion-MNIST: a novel image dataset for benchmarking machine learning algorithms","author":"Xiao","year":"2017"},{"issue":"3","key":"10.7717\/peerj-cs.3461\/ref-41","doi-asserted-by":"publisher","first-page":"158","DOI":"10.1007\/s11063-024-11624-x","article-title":"Class-balanced regularization for long-tailed recognition","volume":"56","author":"Xu","year":"2024","journal-title":"Neural Processing Letters"},{"key":"10.7717\/peerj-cs.3461\/ref-42","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1505.00853","article-title":"Empirical evaluation of rectified activations in convolutional network","author":"Xu","year":"2015"},{"key":"10.7717\/peerj-cs.3461\/ref-43","doi-asserted-by":"publisher","first-page":"137","DOI":"10.1007\/s10462-024-10759-6","article-title":"A survey on imbalanced learning: latest research, applications and future directions","volume":"57","author":"Zhang","year":"2024","journal-title":"Artificial Intelligence Review"},{"issue":"19","key":"10.7717\/peerj-cs.3461\/ref-44","doi-asserted-by":"publisher","first-page":"9071","DOI":"10.3390\/app14199071","article-title":"Imbalanced data parameter optimization of convolutional neural networks based on analysis of variance","volume":"14","author":"Zou","year":"2024","journal-title":"Applied Sciences"}],"container-title":["PeerJ Computer Science"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/peerj.com\/articles\/cs-3461.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3461.xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3461.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3461.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,6]],"date-time":"2026-02-06T08:03:57Z","timestamp":1770365037000},"score":1,"resource":{"primary":{"URL":"https:\/\/peerj.com\/articles\/cs-3461"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,6]]},"references-count":44,"alternative-id":["10.7717\/peerj-cs.3461"],"URL":"https:\/\/doi.org\/10.7717\/peerj-cs.3461","archive":["CLOCKSS","LOCKSS","Portico"],"relation":{},"ISSN":["2376-5992"],"issn-type":[{"value":"2376-5992","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,6]]},"article-number":"e3461"}}