基于迁移类内变化增强数据的小样本学习方法

doi:10.19678/j.issn.1000-3428.0069754

摘要/Abstract

摘要：

小样本学习致力于通过极少数量的训练样本, 甚至一个样本来实现对新类数据的分类。面对这种挑战, 数据增强成为小样本学习中一种直接而有效的解决方法, 但是确保增强数据的多样性和可辨别性是数据增强的关键。为此, 提出一种基于迁移基类类内变化的两阶段数据增强方法, 分为特征学习和小样本学习阶段。在特征学习阶段, 模型通过自监督任务学习基类数据的个体特征表达, 有监督任务则学习类辨别特征, 模型通过这两种特征获得基类数据的类内变化并建模基类的类内变化分布。在小样本学习阶段, 模型从基类的类内变化分布中采样与任务相关的类内变化信息并添加到小样本特征中, 以实现增强小样本数据的目的。实验结果表明, 在5-way 1-shot情况下, 所提方法在miniImageNet、tieredImageNet和CUB数据集上的分类性能相较于基线模型提升了4~7百分点, 在5-way 5-shot情况下提升了3~7百分点, 相较于其他数据增强方法, 也展现了具有竞争力的性能, 这表明生成的增强数据在保持可辨别性的同时增强小样本数据的多样性, 并验证了该方法的可行性和有效性。

关键词: 小样本学习, 数据增强, 类内变化, 类辨别特征, 个体特征

Abstract:

Few-shot learning aims to classify new categories based on only one or a few examples. To address this problem, data augmentation is often used as a direct and effective approach. Further, the augmented data should be diverse and discriminable. This paper proposes a new two-stage data augmentation method based on the transfer of intra-class variations in the base classes. The learning process is divided into a representation learning stage and a few-shot learning stage. In the representation learning stage, representations of instance-specific features of the base-class data are obtained through self-supervised tasks and class-specific features are obtained through supervised tasks. Intra-class variations of the base-class data are calculated using these two features, and the distributions of intra-class variations for each base class are modeled. In the few-shot learning stage, the model samples task-related intra-class variation information from the intra-class variation distributions of the base class and adds it to the few-shot features to enhance the few-shot data. The experimental results show that the 5-way 1-shot classification performance of the proposed method on the miniImageNet, tieredImageNet, and CUB datasets is improved by 4 to 7 percentage points compared to the baseline model, and by 3 to 7 percentage points in 5-way 5-shot classification. Its performance is competitive when compared with other existing data augmentation methods. This indicates that the generated enhanced data can improve the diversity of few-shot data while maintaining discriminability, verifying our method's feasibility and effectiveness.

Key words: few-shot learning, data augmentation, intra-class variation, class-specific feature, instance-specific feature

李小雨, 罗娜. 基于迁移类内变化增强数据的小样本学习方法[J]. 计算机工程, 2025, 51(9): 242-251.

LI Xiaoyu, LUO Na. Few-Shot Learning Method with Augmentation Data Based on Transferring Intra-Class Variations[J]. Computer Engineering, 2025, 51(9): 242-251.

https://www.ecice06.com/CN/Y2025/V51/I9/242

图/表 9

图1 模型学习基类类内变化分布的流程

Fig.1 Procedure of learning the distribution of base class intra-class variation of model

图2 小样本学习阶段的流程

Fig.2 Procedure of few-shot learning stage

图3 在miniImageNet数据集上5-way 1-shot和5-way 5-shot任务的实验结果

Fig.3 Experimental results of 5-way 1-shot and 5-way 5-shot tasks on miniImageNet dataset

图4 特征的t-SNE可视化结果

Fig.4 t-SNE visualization results of features

参考文献 34

1	KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks. Communications of the ACM, 2017, 60(6): 84- 90. doi: 10.1145/3065386
2	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2016: 770-778.
3	陈良臣, 傅德印. 面向小样本数据的机器学习方法研究综述. 计算机工程, 2022, 48(11): 1- 13. doi: 10.19678/j.issn.1000-3428.0065347
	CHEN L C , FU D Y . Survey on machine learning methods for small sample data. Computer Engineering, 2022, 48 (11): 1- 13. doi: 10.19678/j.issn.1000-3428.0065347
4	LI X Z, SUN Q R, LIU Y Y, et al. Learning to self-train for semi-supervised few-shot classification[C]//Proceedings of Conference on Advances in Neural Information Processing Systems. [S. l. ]: AAAI Press, 2019: 10276-10286.
5	AFRASIYABI A, LALONDE J F, GAGNÉ C. Associative alignment for few-shot image classification[C]//Proceedings of the 16th European Conference. Berlin, Germany: Springer, 2020: 18-35.
6	LI J, WANG Z, HU X. Learning intact features by erasing-inpainting for few-shot classification[C]//Proceedings of the AAAI Conference on Artificial Intelligence. [S. l. ]: AAAI Press, 2021: 8401-8409.
7	XU J Y, LE H. Generating representative samples for few-shot classification[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2022: 9003-9013.
8	KINGMA D P, WELLING M. Auto-encoding variational Bayes[EB/OL]. [2024-03-12]. http://export.arxiv.org/pdf/1312.6114.
9	LI K, ZHANG Y L, LI K P, et al. Adversarial feature hallucination networks for few-shot learning[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2020: 13470-13479.
10	GAO H, SHOU Z, ZAREIAN A, et al. Low-shot learning via covariance-preserving adversarial augmentation networks[EB/OL]. [2024-03-12]. https://arxiv.org/pdf/1810.11730.
11	HARIHARAN B, GIRSHICK R. Low-shot visual recognition by shrinking and hallucinating features[C]//Proceedings of IEEE International Conference on Computer Vision (ICCV). Washington D. C., USA: IEEE Press, 2017: 3037-3046.
12	SCHWARTZ E, KARLINSKY L, SHTOK J, et al. Delta-Encoder: an effective sample synthesis method for few-shot object recognition[C]//Proceedings of the 32nd International Conference on Neural Information Processing Systems. New York, USA: ACM Press, 2018: 2850-2860.
13	XU J Y, LE H, HUANG M Z, et al. Variational feature disentangling for fine-grained few-shot classification[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2021: 8812-8821.
14	VINYALS O, BLUNDELL C, LILLICRAP T, et al. Matching networks for one shot learning[C]//Proceedings of the 30th International Conference on Neural Information Processing System. New York, USA: ACM Press, 2016: 3637-3645.
15	REN M, TRIANTAFILLOU E, RAVI S, et al. Meta-learning for semi-supervised few-shot classification[EB/OL]. [2024-03-12]. https://arxiv.org/pdf/1803.00676.
16	WAH C, BRANSON S, WELINDER P, et al. The Caltech-UCSD Birds-200-2011 dataset[EB/OL]. [2024-03-12]. https://gwern.net/doc/ai/dataset/2011-wah.pdf.
17	HE K M, FAN H Q, WU Y X, et al. Momentum contrast for unsupervised visual representation learning[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2020: 9729-9738.
18	OORD A, LI Y Z, VINYALS O. Representation learning with contrastive predictive coding[EB/OL]. [2024-03-12]. https://arxiv.org/pdf/1807.03748.
19	LIN X D, DUAN Y Q, DONG Q Y, et al. Deep variational metric learning[C]//Proceedings of the 15th European Conference on Computer Vision. Berlin, Germany: Springer, 2018: 714-729.
20	YIN X, YU X, SOHN K, et al. Feature transfer learning for deep face recognition with under-represented data[C]//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2019: 233-241.
21	DENG J, DONG W, SOCHER R, et al. ImageNet: a large-scale hierarchical image database[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2009: 248-255.
22	RAVI S, LAROCHELLE H. Optimization as a model for few-shot learning[C]//Proceedigns of the 4th International Conference on Learning Representations. [S. l. ]: AAAI Press, 2016: 120-126.
23	CHEN T, KORNBLITH S, NOROUZI M, et al. A simple framework for contrastive learning of visual representations[EB/OL]. [2024-03-12]. https://arxiv.org/pdf/2002.05709.
24	SNELL J, SWERSKY K, ZEMEL R. Prototypical networks for few-shot learning[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York, USA: ACM Press, 2017: 4080-4090.
25	SUNG F, YANG Y, ZHANG L, et al. Learning to compare: relation network for few-shot learning[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2018: 1199-1208.
26	FINN C, ABBEEL P, LEVINE S. Model-agnostic meta-learning for fast adaptation of deep networks[C]//Proceedings of the 34th International Conference on Machine Learning. New York, USA: ACM Press, 2017: 1126-1135.
27	ZHANG M, HUANG S, WANG D. Domain generalized few-shot image classification via meta regularization network[C]//Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing. Washington D. C., USA: IEEE Press, 2022: 3748-3752.
28	RUSU A A, RAO D, SYGNOWSKI J, et al. Meta-learning with latent embedding optimization[EB/OL]. [2024-03-12]. https://arxiv.org/abs/1807.05960.
29	ORESHKIN B N, RODRIGUEZ P, LACOSTE A. TADAM: task dependent adaptive metric for improved few-shot learning[C]//Proceedings of Advances in Neural Information Processing Systems. [S. l. ]: AAAI Press, 2018: 719-729.
30	JIA J F, FENG X, YU H Q. Few-shot classification via efficient meta-learning with hybrid optimization. Engineering Applications of Artificial Intelligence, 2024, 127, 107296. doi: 10.1016/j.engappai.2023.107296
31	CHEN J, HU Y, SHEN M, et al. Dual Episodic sampling and momentum consistency regularization for unsupervised few-shot learning[C]//Proceedings of IEEE International Conference on Multimedia and Expo (ICME). Washington D. C., USA: IEEE Press, 2023: 2891-2896.
32	ZHANG L, ZHOU F, WEI W, et al. Meta-hallucinating prototype for few-shot learning promotion. Pattern Recognition: The Journal of the Pattern Recognition Society, 2023, 136, 109235. doi: 10.1016/j.patcog.2022.109235
33	LEE K, MAJI S, RAVICHANDRAN A, et al. Meta-learning with differentiable convex optimization[C]//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2019: 10657-10665.
34	SUN Q, LIU Y, CHUA T S, et al. Meta-transfer learning for few-shot learning[C]//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Washington D. C., USA: IEEE Press, 2019: 403-412.

[1]	马淦, 谷雨, 彭冬亮. 结合改进YOLOv5s和动态数据增强的海面舰船检测[J]. 计算机工程, 2025, 51(9): 294-305.
[2]	朱思远, 李佳圣, 邹丹平, 何迪, 郁文贤. 基于半监督学习的非结构化道路缺陷检测算法[J]. 计算机工程, 2025, 51(9): 14-24.
[3]	王帅, 史艳翠. 基于个性化数据增强的自监督序列推荐算法[J]. 计算机工程, 2025, 51(8): 190-202.
[4]	庞鑫, 葛凤培, 李艳玲. 声景识音：数字化时代声学场景分类的探索与前沿[J]. 计算机工程, 2025, 51(6): 1-19.
[5]	商雅名, 吴安彪, 袁野, 王一舒. 基于个性化PageRank高阶邻域聚合的图神经网络增强[J]. 计算机工程, 2025, 51(6): 38-48.
[6]	张河萍, 方志军, 卢俊鑫, 高永彬. 基于知识增强自适应原型网络的小样本关系分类[J]. 计算机工程, 2025, 51(4): 129-136.
[7]	张兴鹏, 何东, 杨模, 叶杭滨. 基于多尺度注意力和数据增强的细胞核分割[J]. 计算机工程, 2025, 51(2): 387-396.
[8]	王月昊, 周若华. 低资源环境下的语音唤醒研究综述[J]. 计算机工程, 2025, 51(2): 35-53.
[9]	李维刚, 厉许昌, 田志强, 李金灵. 基于自蒸馏框架的点云分类及其鲁棒性研究[J]. 计算机工程, 2024, 50(9): 72-81.
[10]	张溢文, 蔡满春, 陈咏豪, 朱懿, 姚利峰. 融合空间特征的多尺度深度伪造检测方法[J]. 计算机工程, 2024, 50(7): 240-250.
[11]	林芷薇, 杨祖元, 王斯秋, 杨超. 基于多尺度线性全局注意力的运动员检测算法[J]. 计算机工程, 2024, 50(7): 352-359.
[12]	张宝鑫, 杨丹, 聂铁铮, 寇月. 基于自监督的多视角图协同过滤推荐方法[J]. 计算机工程, 2024, 50(5): 100-110.
[13]	宫阿娟, 潘天荣. 多病种眼底疾病诊断的深度学习策略讨论[J]. 计算机工程, 2024, 50(5): 363-372.
[14]	侯钰涛, 阿布都克力木·阿布力孜, 史亚庆, 马依拉木·木斯得克, 哈里旦木·阿布都克里木. 面向"一带一路"的低资源语言机器翻译研究[J]. 计算机工程, 2024, 50(4): 332-341.
[15]	司明悦, 齐斌, 张文胜, 张雷. 基于张量计算的智慧交通多维数据计算与小样本学习[J]. 计算机工程, 2024, 50(4): 41-49.

选择文件类型/文献管理软件名称

选择包含的内容