Imbalanced Image Classification Method Based on Kernel Feature Scaling and Boundary Sample Mining

doi:10.3778/j.issn.1673-9418.2408078

Abstract

Abstract: Imbalanced image classification remains a formidable challenge in computer vision, where deep learning has been extensively applied but still struggles with class imbalance, causing models to bias towards the majority class and overlook the minority class. Traditional data sampling techniques can introduce noise or omit essential information, thus impeding model generalization. Integrating geometric concepts into deep learning has emerged as an effective solution, improving the classification performance by restructuring feature space, refining decision boundaries, and diversifying data.This paper introduces a novel geometric deep learning method that combines kernel-based hierarchical feature scaling (KHFS) with relative neighborhood boundary mining (RNBM). Inspired by support vector data description (SVDD), KHFS utilizes kernel functions to hierarchically cluster and identify class centroids, scaling feature vectors based on the radius of hyper-spheres centered at these centroids to enhance the representation of minority class samples. RNBM constructs a relative neighborhood graph to identify boundary samples at class intersections, promoting intra-class compactness and inter-class separation. Furthermore, the convolutional block self-attention (CBSA) mechanism is integrated into the CNN feature extraction module to concentrate on salient input features. Comprehensive experiments on CIFAR-10, CIFAR-100 and CINIC-10 datasets demonstrate the superior performance of proposed geometric deep learning method in mitigating data imbalance compared with existing models.

Key words: imbalanced image classification, geometric methods, feature scaling, boundary sample mining, support vector data description

摘要： 不平衡图像分类是计算机视觉领域的一大挑战。尽管深度学习技术已被广泛应用，但类别不平衡问题仍然显著，导致模型偏向多数类而忽视少数类。传统的数据采样方法易引入噪声或丢失关键信息，限制了模型的泛化能力。研究表明，将几何思想融入深度学习方法中是一种有效且创新的解决方案。几何思想通过优化特征空间结构、改进决策边界和增强数据多样性，显著提升了不平衡图像分类的性能。提出了一种新的几何深度学习方法，该方法集成了基于核函数的层次特征缩放技术（KHFS）和相对邻域边界样本挖掘手段（RNBM）。KHFS借鉴了基于核函数的支持向量数据描述（SVDD），通过层次聚类确定每类的中心点，并计算以中心点为球心的超球体半径，对各类特征向量进行相应缩放，从而增强少数类样本的表示能力。RNBM方法则通过构建相对邻域图来捕捉样本间的邻域关系，从中挖掘出不同类别交界处的边界样本，以约束类内样本的紧凑性和类间样本的分散性。此外，也引入了卷积块自注意力机制（CBSA），应用于卷积神经网络（CNN）特征提取模块，旨在关注输入数据中的关键信息。在CIFAR-10、CIFAR-100、CINIC-10三个基准数据集上的大量实验验证了该几何深度学习方法在解决数据不平衡问题方面优于现有模型的显著性能。

关键词: 不平衡图像分类, 几何方法, 特征缩放, 边界样本挖掘, 支持向量数据描述

LENG Qiangkui, TAO Shuqing. Imbalanced Image Classification Method Based on Kernel Feature Scaling and Boundary Sample Mining[J]. Journal of Frontiers of Computer Science and Technology, 2025, 19(7): 1851-1867.

冷强奎, 陶抒清. 基于核特征缩放和边界样本挖掘的不平衡图像分类方法[J]. 计算机科学与探索, 2025, 19(7): 1851-1867.

References

[1] TRAN K A, KONDRASHOVA O, BRADLEY A, et al. Deep learning in cancer diagnosis, prognosis and treatment selection[J]. Genome Medicine, 2021, 13(1): 152.
[2] AGGARWAL C C, YU P S. Outlier detection for high dimensional data[C]//Proceedings of the 2001 ACM SIGMOD International Conference on Management of Data. New York: ACM, 2001: 37-46.
[3] BUDA M, MAKI A, MAZUROWSKI M A. A systematic study of the class imbalance problem in convolutional neural networks[J]. Neural Networks, 2018, 106: 249-259.
[4] ZHANG Y F, KANG B Y, HOOI B, et al. Deep long-tailed learning: a survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(9): 10795-10816.
[5] HE H B, GARCIA E A. Learning from imbalanced data[J]. IEEE Transactions on Knowledge and Data Engineering, 2009, 21(9): 1263-1284.
[6] WANG Z, ZHU Z H, LI D D. Collaborative and geometric multi-kernel learning for multi-class classification[J]. Pattern Recognition, 2020, 99: 107050.
[7] ZHU Y J, WANG Z, ZHA H Y, et al. Boundary-eliminated pseudoinverse linear discriminant for imbalanced problems[J]. IEEE Transactions on Neural Networks and Learning Systems, 2018, 29(6): 2581-2594.
[8] GHAZIKHANI A, YAZDI H S, MONSEFI R. Class imbalance handling using wrapper-based random oversampling[C]//Proceedings of the 20th Iranian Conference on Electrical Engineering. Piscataway: IEEE, 2012: 611-616.
[9] 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.
[10] HAN H, WANG W Y, MAO B H. Borderline-SMOTE: a new over-sampling method in imbalanced data sets learning[C]//Proceedings of the 2005 International Conference on Intelligent Computing. Berlin, Heidelberg: Springer, 2005: 878-887.
[11] HE H B, BAI Y, GARCIA E A, et al. ADASYN: adaptive synthetic sampling approach for imbalanced learning[C]//Proceedings of the 2008 IEEE International Joint Conference on Neural Networks. Piscataway: IEEE, 2008: 1322-1328.
[12] GARCíA V, SáNCHEZ J S, MARQUéS A I, et al. Understanding the apparent superiority of over-sampling through an analysis of local information for class-imbalanced data[J]. Expert Systems with Applications, 2020, 158: 113026.
[13] YEN S J, LEE Y S. Cluster-based under-sampling approaches for imbalanced data distributions[J]. Expert Systems with Applications, 2009, 36(3): 5718-5727.
[14] LIU X Y, WU J X, ZHOU Z H. Exploratory undersampling for class-imbalance learning[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2009, 39(2): 539-550.
[15] KUKAR M, KONONENKO I. Cost-sensitive learning with neural networks[C]//Proceedings of the 13th European Conference on Artificial Intelligence. Hoboken: John Wiley & Sons, 1998: 88-94.
[16] SUN Y M, KAMEL M S, WONG A K C, et al. Cost-sensitive boosting for classification of imbalanced data[J]. Pattern Recognition, 2007, 40(12): 3358-3378.
[17] WONG M L, SENG K, WONG P K. Cost-sensitive ensemble of stacked denoising autoencoders for class imbalance problems in business domain[J]. Expert Systems with Applications, 2020, 141: 112918.
[18] GENG Y, LUO X Y. Cost-sensitive convolutional neural networks for imbalanced time series classification[J]. Intelligent Data Analysis, 2019, 23(2): 357-370.
[19] CHUNG Y A, YANG S W, LIN H T. Cost-sensitive deep learning with layer-wise cost estimation[C]//Proceedings of the 2020 International Conference on Technologies and Applications of Artificial Intelligence. Piscataway: IEEE, 2020: 108-113.
[20] LU Y, CHEUNG Y M, TANG Y Y. Hybrid sampling with bagging for class imbalance learning[C]//Proceedings of the 2016 Pacific-Asia Conference on Knowledge Discovery and Data Mining. Cham: Springer, 2016: 14-26.
[21] CHAWLA N V, LAZAREVIC A, HALL L O, et al. SMOTEBoost: improving prediction of the minority class in boosting[C]//Proceedings of the 7th European Conference on Principles and Practice of Knowledge Discovery in Databases. Berlin, Heidelberg: Springer, 2003: 107-119.
[22] SEIFFERT C, KHOSHGOFTAAR T M, VAN HULSE J, et al. RUSBoost: a hybrid approach to alleviating class imbalance[J]. IEEE Transactions on Systems, Man, and Cybernetics - Part A: Systems and Humans, 2010, 40(1): 185-197.
[23] FAN W, STOLFO S J, ZHANG J, et al. AdaCost: misclassification cost-sensitive boosting[C]//Proceedings of the 16th International Conference on Machine Learning. San Francisco: Morgan Kaufmann, 1999: 97-105.
[24] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2017: 2999-3007.
[25] CAO K D, WEI C, GAIDON A, et al. Learning imbalanced datasets with label-distribution-aware margin loss[EB/OL]. [2024-06-17]. https://arxiv.org/abs/1906.07413.
[26] ZHOU B Y, CUI Q, WEI X S, et al. BBN: bilateral-branch network with cumulative learning for long-tailed visual recognition[C]//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 9716-9725.
[27] CUI Y, JIA M L, LIN T Y, et al. Class-balanced loss based on effective number of samples[C]//Proceedings of the 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2019: 9260-9269.
[28] BRONSTEIN M M, BRUNA J, COHEN T, et al. Geometric deep learning: grids, groups, graphs, geodesics, and gauges [EB/OL]. [2024-06-17]. https://arxiv.org/abs/2104.13478.
[29] MONTI F, BOSCAINI D, MASCI J, et al. Geometric deep learning on graphs and manifolds using mixture model CNNs[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 5425-5434.
[30] BRONSTEIN M M, BRUNA J, LECUN Y, et al. Geometric deep learning: going beyond Euclidean data[J]. IEEE Signal Processing Magazine, 2017, 34(4): 18-42.
[31] DONG C, LOY C C, HE K M, et al. Image super-resolution using deep convolutional networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(2): 295-307.
[32] GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al. Generative adversarial networks[EB/OL]. [2024-06-17]. https:// arxiv.org/abs/1406.2661.
[33] PUJOL O, MASIP D. Geometry-based ensembles: toward a structural characterization of the classification boundary[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(6): 1140-1146.
[34] ZHU Z H, WANG Z, LI D D, et al. Geometric structural ensemble learning for imbalanced problems[J]. IEEE Transactions on Cybernetics, 2020, 50(4): 1617-1629.
[35] TORRES L C B, CASTRO C L, COELHO F, et al. Large margin Gaussian mixture classifier with a Gabriel graph geometric representation of data set structure[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 32(3): 1400-1406.
[36] GHASEMIGOL M, MONSEFI R, YAZDI H S. Ellipse support vector data description[C]//Proceedings of the 11th Inter-national Conference on Engineering Applications of Neural Networks. Berlin, Heidelberg: Springer, 2009: 257-268.
[37] TAX D M J, DUIN R P W. Support vector data description[J]. Machine Learning, 2004, 54(1): 45-66.
[38] KAKDE D, CHAUDHURI A, KONG S, et al. Peak criterion for choosing Gaussian kernel bandwidth in support vector data description[C]//Proceedings of the 2017 IEEE International Conference on Prognostics and Health Management. Piscataway: IEEE, 2017: 32-39.
[39] ZHOU Z H, LIU X Y. Training cost-sensitive neural networks with methods addressing the class imbalance problem[J]. IEEE Transactions on Knowledge and Data Engineering, 2006, 18(1): 63-77.
[40] KIM M J, KANG D K, KIM H B. Geometric mean based boosting algorithm with over-sampling to resolve data imbalance problem for bankruptcy prediction[J]. Expert Systems with Applications, 2015, 42(3): 1074-1082.
[41] ZHANG S C, LI X L, ZONG M, et al. Efficient kNN classification with different numbers of nearest neighbors[J]. IEEE Transactions on Neural Networks and Learning Systems, 2018, 29(5): 1774-1785.
[42] WARD JR J H. Hierarchical grouping to optimize an objective function[J]. Journal of the American Statistical Association, 1963, 58(301): 236-244.
[43] WANG Z, DONG Q D, GUO W, et al. Geometric imbalanced deep learning with feature scaling and boundary sample mining[J]. Pattern Recognition, 2022, 126: 108564.
[44] NAPIERA?A K, STEFANOWSKI J, WILK S. Learning from imbalanced data in presence of noisy and borderline examples[C]//Proceedings of the 7th International Conference on Rough Sets and Current Trends in Computing. Berlin, Heidelberg: Springer, 2010: 158-167.
[45] DONG Q, GONG S G, ZHU X T. Class rectification hard mining for imbalanced deep learning[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2017: 1869-1878.
[46] HOANG C, CHOI Y, CARLBERG K. Domain-decomposition least-squares Petrov-Galerkin (DD-LSPG) nonlinear model reduction[J]. Computer Methods in Applied Mechanics and Engineering, 2021, 384: 113997.
[47] WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]//Proceedings of the 15th European Conference on Computer Vision. Cham: Springer, 2018: 3-19.
[48]VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[C]//Advances in Neural Information Processing Systems 30, 2017: 5998-6008.
[49] NAM H, HA J W, KIM J. Dual attention networks for multimodal reasoning and matching[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 2156-2164.
[50] KRIZHEVSKY A, HINTON G. Learning multiple layers of features from tiny images[EB/OL]. [2024-06-17].https://scholar.google.com.hk/scholar?hl=zh-CN&as_sdt=0%2C5&q=Learning+ multiple+layers+of+features+from+tiny+images&btnG=.
[51] DARLOW L N, CROWLEY E J, ANTONIOU A, et al. CINIC-10 is not ImageNet or CIFAR-10[EB/OL]. [2024-06-20]. https://arxiv.org/abs/1810.03505.
[52] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 7132-7141.
[53] WANG F, JIANG M Q, QIAN C, et al. Residual attention network for image classification[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 6450-6458.