基于聚类优化学习的少样本图像分类

doi:10.11896/jsjkx.230300227

摘要/Abstract

摘要： 少样本图像分类的目标是在训练少量标记训练数据的基础上实现新类别图像的分类,然而这一目的在现有条件下很难实现。因此,目前的少样本学习方法主要借鉴迁移学习的思想,其核心是利用情景训练式的元训练构建先验知识,从而解决未知新任务。然而,研究工作表明,相较于复杂的少样本学习方法,具有强大特征表示的嵌入模型学习方法更为简单、有效。受此启发,提出一种新的基于直推式聚类优化学习的少样本图像分类方法。该方法首先利用样本数据的内部特征结构信息实现每个类别的综合表示;然后优化每个类别的中心,形成更具区别性的特征表示,从而有效增加不同类别之间的特征差异。大量实验结果表明,所提的基于直推式聚类优化学习的少镜头图像分类方法有效提高了各种训练条件下的图像分类精度。

关键词: 少样本图像分类, 特征表示, 聚类优化

Abstract: The goal of few-shot image classificationis to achieve the classification of new imagecategories on the basis of training a small number of labeled training dataset.However,this goal is difficult to achieve under existing conditions.Therefore,the current few-shot learning method mainly mainly draws on the idea of transfer learning,and its core is to construct prior knowledge by using situational meta-training,so as to realize the solution of unknown new tasks.However,the latest research shows that the embedded model learning method with strong feature representation is simpler and more effective than the complex few-shot learning method.Inspired by this,this paper proposes a novel few-shot image classification methodbased on direct clustering optimization learning.This proposed method first utilizes the internal feature structure information of sample data to realize the comprehensive representation of each category,and then optimizes the center of each category to form a more distinctive feature representation,thus effectively increasing the feature differences between different categories.A large number of experimental results demonstrate that the proposed image classification method based on the clustering optimization learningcan effectively improve the accuracy of image classification under various training conditions.

Key words: Few-shot image classification, Feature representation, Clustering optimization

中图分类号:

TP391.41

苏如祺, 卞雄, 朱松豪. 基于聚类优化学习的少样本图像分类[J]. 计算机科学, 2024, 51(6A): 230300227-7. https://doi.org/10.11896/jsjkx.230300227

SU Ruqi, BIAN Xiong, ZHU Songhao. Few-shot Images Classification Based on Clustering Optimization Learning[J]. Computer Science, 2024, 51(6A): 230300227-7. https://doi.org/10.11896/jsjkx.230300227

参考文献

[1]HE K M,GKIOXARI G,DOLL′AR P,et al.Mask R-CNN[C]//IEEE Conference on Computer Vision,2017:2961-2969.
[2]SU X,HUANG T,LI Y L,et al.Prioritized architecture sampling with monto-carlo tree search[C]//IEEE Conference on Computer Vision and Pattern Recognition.2021:10968-10977.
[3]RUSU A A,RAO D,SYGNOWSKI J,et al.Meta-learning with latent embedding optimization[C]//International Conference on Learning Representations.2019:1-17.
[4]BERTINETTO L,HENRIQUES J F,TORR P H S,et al.Meta-learning with differentiable closed-form solvers[C]//International Conference on Learning Representations.2019:1-15.
[5]LAENEN S,BERTINETTO L.On episodes,prototypical net-works,and few-shot learning[C]//Annual Conference on Neural Information Processing Systems.2021:24581-24592.
[6]DHILLON G S,CHAUDHARI P,RAVICHANDRAN A,et al.A baseline for few-shot image classification[C]//International Conference on Learning Representations.2020:1-20.
[7]TIAN Y L,WANG Y,KRISHNAN D,et al.Rethinking few-shot image classification:a good embedding is all you need?[C]//European Conference on Computer Vision.2020:266-282.
[8]CHEN W Y,LIU Y C,KIRA Z,et al.A closer look at few-shot classification[C]//International Conference on Learning Representations.2019:1-16.
[9]WANG Y,CHAO W L,WEINBERGER K Q,et al.SimpleShot:Revisiting Nearest-Neighbor Classification for Few-Shot Lear-ning[J].arXiv:1911.04623,2019.
[10]YANG S,LIU L,XU M.Free lunch for few-shot learning:Distribution calibration[C]//International Conference on Learning Representations.2021:1-13.
[11]LUO X,WEI L H,WEN L J,et al.Rectifying the shortcutlearning of background for few-shot learning[C]//Annual Conference on Neural Information Processing Systems.2021:13073-13085.
[12]RODRÍGUEZ P,LARADJI I H,DROUIN A,et al.Embedding propagation:Smoother manifold for few-shot classification[C]//European Conference on Computer Vision.2020:121-138.
[13]SNELL J,SWERSKY K,ZEMEL R S.Prototypical networksfor few-shot learning[C]//Annual Conference on Neural Information Processing Systems.2017:4077-4087.
[14]SUNG F,YANG Y X,ZHANG L,et al.Learning to compare:Relation network for few-shot learning[C]//IEEE Conference on Computer Vision and Pattern Recognition.2018:1199-1208.
[15]LI W B,WANG L,XU J L,et al.Revisiting local descriptorbased image-to-class measure for few-shot learning[C]//IEEE Conference on Computer Vision and Pattern Recognition.2019:7260-7268.
[16]ZHANG C,CAI Y J,LIN G S,et al.DeepEMD:Few-shot image classification with differentiable earth mover’s distance and structured classifiers[C]//IEEE Conference on Computer Vision and Pattern Recognition.2020:12200-12210.
[17]FINN C,ABBEEL P,LEVINE S.Modelagnostic meta-learningfor fast adaptation of deep networks[C]//International Conference on Machine Learning.2017:1126-1135.
[18]NICHOL A,ACHIAM J,SCHULMAN J.On first-order meta-learning algorithms[J].arXiv:1803.02999,2018.
[19]TRIANTAFIFILLOU E,ZHU T,DUMOULIN V,et al.Meta-dataset:A dataset of datasets for learning to learn from few examples[C]//International Conference on Learning Representations.2020:1-24.
[20]WU Z R,XIONG Y J,YU S X,et al.Unsupervised Feature Learning via Non-Parametric Instance-level Discrimination[J].arXiv:1805.0197,2018.
[21]YE M,ZHANG X,YUEN P C,et al.Unsupervised EmbeddingLearning via Invariant and Spreading Instance Feature[C]//IEEE Conference on Computer Vision and Pattern Recognition.2019:6210-6219.
[22]CHEN T,KORNBLITH S,NOROUZI M,et al.A SimpleFramework for Contrastive Learning of Visual Representations[C]//International Conference on Learning Representations.2020:1597-1607.
[23]LIU C,ZHANG L,XU C M,et al.Learning a Few-shot Embedding Model with Contrastive Learning[C]//AAAI Conference on Artificial Intelligence.2021:8635-8643.
[24]LUO X,CHEN Y X,WEN L J,et al.Boosting few-shot classification with view-learnable contrastive learning[C]//IEEE Conference on Multimedia and Expo.2021:1-6.
[25]MIYATO T,MAEDA S I,KOYAMA M,et al.Virtual adver-sarial training:a regularization method for supervised and semi-supervised learning[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2018,41(8):1979-1993.
[26]WEI C,SOHN K,MELLINA C,et al.Crest:A class-rebalancing self-training framework for imbalanced semi-supervised learning[C]//IEEE Conference on Computer Vision and Pattern Recognition.2021:10857-10866.
[27]REN M Y,TRIANTAFILLOU E,RAVI S,et al.Meta-learning for semi-supervised few-shot classification[C]//International Conference on Learning Representations.2018:1-15.
[28]HU Y Q,GRIPON V,PATEUX S.Graph-based interpolation of feature vectors for accurate few-shot classification[C]//International Conference on Pattern Recognition.2021:8164-8171.
[29]PAPYAN V,HAN X Y,DONOHO D L.Prevalence of neural collapse during the terminal phase of deep learning training[J].arXiv:2008.08186,2020.
[30]VINYALS O,BLUNDELL C,LILLICRAP T,et al.Matchingnetworks for one shot learning[C]//Annual Conference on Neural Information Processing Systems.2016:3630-3638.
[31]ORESHKIN B,RODR′IGUEZ L′OPEZ P,LACOSTE A.Ta-dam:Task dependent adaptive metric for improved few-shot learning[C]//Annual Conference on Neural Information Processing Systems.2018:721-731.
[32]RAVI S,LAROCHELLE H.Optimization as a model for fewshot learning[C]//International Conference on Learning Representations.2017:1-11.
[33]MISHRA N,ROHANINEJAD M,CHEN X,et al.A simpleneural attentive meta-learner[C]//International Conference on Learning Representations.2018:1-17.
[34]LEE K,MAJI S,RAVICHANDRAN A,et al.Meta-learningwith differentiable convex optimization[C]//IEEE Conference on Computer Vision and Pattern Recognition.2019:10657-10665.
[35]HILLER M,MA R K,HARANDI M,et al.Rethinking generalization in few-shot classification[J].arXiv:2206.07267,2022.
[36]YANG L,LI L L,ZHANG Z L,et al.DPGN:Distribution prop-agation graph network for few-shot learning[C]//IEEE Conference on Computer Vision and Pattern Recognition.2020:13387-13396.
[37]SHEN X,XIAO Y,HU S X,et al.Re-ranking for image retrievaland transductive few-shot classification[C]//Annual Conference on Neural Information Processing Systems.2021:25932-25943.
[38]LAZAROU M,STATHAKI T,AVRITHIS Y.Iterative labelcleaning for transductive and semi-supervised few-shot learning[C]//IEEE Conference on Computer Vision.2021:8731-8740.
[39]CHEN C F,YANG X S,XU C S,et al.ECKPN:Explicit class knowledge propagation network for transductive few-shot learning[C]//IEEE Conference on Computer Vision and Pattern Recognition.2021:6592-6601.
[40]BENDOU Y,HU Y Q,LAFARGUE R,et al.EASY:Ensemble augmented-shot y-shaped learning:state-of-the-art few-shot classification with simple ingredients[J].arXiv:2201.09699,2022.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed