A few-shot image classification method based on Ball k-means algorithm
DOI:
https://doi.org/10.54097/7pgzq265Keywords:
Ball k-means, k-means, Iteration, Image classification, iAdaptabilityAbstract
Ball K-means algorithm is an accelerated and precise clustering method that uses spheres to represent clusters, thereby reducing the distance calculations between data points and centroids. Initially designed for graph data, it was not efficient or accurate for clustering centroids in image data. This paper adjusts the ball K-means algorithm and integrates it with image analysis, expanding its application domain and introducing a novel theoretical framework. The method represents each class of query samples using spherical clusters, identifies neighboring clusters, and distinguishes between active and stable regions. It adaptively updates centroids based on the correlation between active region sample points and neighboring cluster centroids. The resulting centroids are combined with class representatives from training samples to form more distinctive spherical prototypes. The Euclidean distance between query samples and these prototypes is computed for classification based on the nearest-neighbor assignment principle, enhancing model generalization through adaptive features. Experimental results demonstrate that our method achieves improved classification accuracy and robustness on miniImageNet and tieredImageNet datasets.
Downloads
References
[1] Ahmed, M., Seraj, R., & Islam, S. M. S. (2020). The k-means algorithm: A comprehensive survey and performance evaluation. Electronics, 9(8), 1295. https://doi.org/10.3390/electronics9081295
[2] Chavan, M., Patil, A., Dalvi, L., & Patil, A. (2015). Mini batch k-means clustering on large dataset. International Journal of Science, Engineering and Technology Research, 4(7), 1356-1358.
[3] Nie, F., Li, Z., Wang, R., & Li, X. (2023). An effective and efficient algorithm for k-means clustering with new formulation. IEEE Transactions on Knowledge and Data Engineering, 35(4), 3433-3443. https://doi.org/10.1109/TKDE.2022.3142576
[4] Zhu, X., Sun, J., He, Z., Jiang, J., & Wang, Z. (2023). Staleness-reduction mini-batch k-means. IEEE Transactions on Neural Networks and Learning Systems, 34(12), 10247-10258. https://doi.org/10.1109/TNNLS.2023.3262814
[5] Xia, S., Peng, D., Meng, D., Zhang, C., Wang, G., Giem, E., Wei, W., & Chen, Z. (2022). Ball k-means: Fast adaptive clustering with no bounds. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(1), 87-99. https://doi.org/10.1109/TPAMI.2020.3038692
[6] Xia, S., Liu, Y., Ding, X., Wang, G., Yu, H., & Luo, Y. (2019). Granular ball computing classifiers for efficient, scalable and robust learning. Information Sciences, 483, 136-152. https://doi.org/10.1016/j.ins.2019.01.010
[7] Xia, S., Zheng, S., Wang, G., Gao, X., & Wang, B. (2023). Granular ball sampling for noisy label classification or imbalanced classification. IEEE Transactions on Neural Networks and Learning Systems, 34(4), 2144-2155. https://doi.org/10.1109/TNNLS.2021.3111362
[8] Xia, S., Dai, X., Wang, G., Gao, X., & Giem, E. (2022). An efficient and adaptive granular-ball generation method in classification problem. IEEE Transactions on Neural Networks and Learning Systems, 33(10), 5610-5624. https://doi.org/10.1109/TNNLS.2022.3161503
[9] Xie, Q., Zhang, Q., Xia, S., Zhao, F., Wu, C., Wang, G., & Ding, W. (2024). GBG++: A fast and stable granular ball generation method for classification. IEEE Transactions on Emerging Topics in Computational Intelligence, 8(3), 2255-2268. https://doi.org/10.1109/TETCI.2024.3372653
[10] Song, Y., Wang, T., Cai, P., Mondal, S. K., & Sahoo, J. P. (2023). A comprehensive survey of few-shot learning: Evolution, applications, challenges, and opportunities. ACM Computing Surveys, 55(13s), Article 271. https://doi.org/10.1145/3582688
[11] Wang, J., Liu, K., Zhang, Y., Leng, B., & Lu, J. (2023). Recent advances of few-shot learning methods and applications. Science China Technological Sciences, 66(4), 920-944. https://doi.org/10.1007/s11431-022-2323-6
[12] Xia, S., Wang, G., & Gao, X. (2023). Granular ball computing: An efficient, robust, and interpretable adaptive multi-granularity representation and computation method. arXiv, arXiv:2304.11171. https://doi.org/10.48550/arXiv.2304.11171
[13] Xia, S., Wu, S., Chen, X., Wang, G., Gao, X., Zhang, Q., Giem, E., & Chen, Z. (2022). GRRS: Accurate and efficient neighborhood rough set for feature selection. IEEE Transactions on Knowledge and Data Engineering, 34(11), 5427-5440. https://doi.org/10.1109/TKDE.2022.3166428
[14] Xia, S., Wang, C., Wang, G., Gao, X., Ding, W., Yu, J., Zhai, Y., & Chen, Z. (2023). GBRS: A unified granular-ball learning model of Pawlak rough set and neighborhood rough set. IEEE Transactions on Neural Networks and Learning Systems, 34(9), 6298-6311. https://doi.org/10.1109/TNNLS.2022.3172692
[15] Li, X., Yang, X., Ma, Z., & Xue, J.-H. (2023). Deep metric learning for few-shot image classification: A review of recent developments. Pattern Recognition, 138, 109381. https://doi.org/10.1016/j.patcog.2023.109381
[16] Wang, Y., Yao, Q., Kwok, J. T., & Ni, L. M. (2020). Generalizing from a few examples: A survey on few-shot learning. ACM Computing Surveys, 53(3), Article 63. https://doi.org/10.1145/3386252
[17] Bellet, A., Habrard, A., & Sebban, M. (2022). Metric learning. Springer Nature. https://doi.org/10.1007/978-3-031-05964-1
[18] Vinyals, O., Blundell, C., Lillicrap, T., Wierstra, D., & Kavukcuoglu, K. (2016). Matching networks for one shot learning. Advances in Neural Information Processing Systems, 29, 3637-3645.
[19] Snell, J., Swersky, K., & Zemel, R. (2017). Prototypical networks for few-shot learning. Advances in Neural Information Processing Systems, 30, 4077-4087.
[20] Sung, F., Yang, Y., Zhang, L., Xiang, T., Torr, P. H. S., & Hospedales, T. M. (2018). Learning to compare: Relation network for few-shot learning. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 1199-1208). IEEE. https://doi.org/10.1109/CVPR.2018.00131
[21] Li, W., Wang, L., Xu, J., Huo, J., Gao, Y., & Luo, J. (2019). Revisiting local descriptor based image-to-class measure for few-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 7260-7268). IEEE. https://doi.org/10.1109/CVPR.2019.00743
[22] Jia, X., Su, Y., & Zhao, H. (2023). Few-shot learning via relation network based on coarse-grained granulation. Applied Intelligence, 53(1), 996-1008. https://doi.org/10.1007/s10489-022-03650-y
[23] Jia, X., Mao, Y., Pan, Z., Wang, Z., & Ping, P. (2024). Few-shot learning based on hierarchical feature fusion via relation networks. International Journal of Approximate Reasoning, 170, 109186. https://doi.org/10.1016/j.ijar.2024.109186
[24] Vanschoren, J. (2019). Meta-learning. In Automated machine learning: Methods, systems, challenges (pp. 35-61). Springer. https://doi.org/10.1007/978-3-030-05318-5_2
[25] Wang, J. X. (2021). Meta-learning in natural and artificial intelligence. Current Opinion in Behavioral Sciences, 38, 90-95. https://doi.org/10.1016/j.cobeha.2020.10.010
[26] Finn, C., Abbeel, P., & Levine, S. (2017). Model-agnostic meta-learning for fast adaptation of deep networks. In International Conference on Machine Learning (pp. 1126-1135). PMLR.
[27] Nichol, A., Achiam, J., & Schulman, J. (2018). On first-order meta-learning algorithms. arXiv, arXiv:1803.02999. https://doi.org/10.48550/arXiv.1803.02999
[28] Antoniou, A., Edwards, H., & Storkey, A. (2019). How to train your MAML. In Seventh International Conference on Learning Representations (ICLR 2019).
[29] Jamal, M. A., & Qi, G.-J. (2019). Task agnostic meta-learning for few-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 11719-11727). IEEE. https://doi.org/10.1109/CVPR.2019.01199
[30] Rajeswaran, A., Finn, C., Kakade, S. M., & Levine, S. (2019). Meta-learning with implicit gradients. Advances in Neural Information Processing Systems, 32, 113-124.
[31] Rajasegaran, J., Khan, S., Hayat, M., Khan, F. S., & Shah, M. (2020). iTAML: An incremental task-agnostic meta-learning approach. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 13588-13597). IEEE. https://doi.org/10.1109/CVPR42600.2020.01360
[32] Kaya, M., & Bilge, H. Ş. (2019). Deep metric learning: A survey. Symmetry, 11(9), 1066. https://doi.org/10.3390/sym11091066
[33] Hospedales, T., Antoniou, A., Micaelli, P., & Storkey, A. (2022). Meta-learning in neural networks: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(9), 5149-5169. https://doi.org/10.1109/TPAMI.2021.3079209
[34] Ren, M., Triantafillou, E., Ravi, S., Snell, J., Swersky, K., Tenenbaum, J. B., Larochelle, H., & Zemel, R. S. (2018). Meta-learning for semi-supervised few-shot classification. In 6th International Conference on Learning Representations (ICLR 2018).
[35] Ye, H.-J., Hu, H., Zhan, D.-C., & Sha, F. (2020). Few-shot learning via embedding adaptation with set-to-set functions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 8808-8817). IEEE. https://doi.org/10.1109/CVPR42600.2020.00883
[36] Patacchiola, M., Turner, J., Crowley, E. J., O’Boyle, M., & Storkey, A. J. (2020). Bayesian meta-learning for the few-shot setting via deep kernels. Advances in Neural Information Processing Systems, 33, 16108-16118.
[37] Oh, J., Yoo, H., Kim, C., & Yun, S.-Y. (2020). BOIL: Towards representation change for few-shot learning. arXiv, arXiv:2008.07408. https://doi.org/10.48550/arXiv.2008.07408
[38] Ravichandran, A., Bhotika, R., & Soatto, S. (2019). Few-shot learning with embedded class models and shot-free meta training. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 331-339). IEEE. https://doi.org/10.1109/ICCV.2019.00042
[39] Huang, H., Wu, Z., Li, W., Huo, J., & Gao, Y. (2021). Local descriptor-based multi-prototype network for few-shot learning. Pattern Recognition, 116, 107935. https://doi.org/10.1016/j.patcog.2021.107935
[40] He, J., Hong, R., Liu, X., Xu, M., & Sun, Q. (2022). Revisiting local descriptor for improved few-shot classification. ACM Transactions on Multimedia Computing, Communications and Applications, 18(2s), Article 53. https://doi.org/10.1145/3472815
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Journal of Computing and Electronic Information Management
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
