计算机科学 ›› 2024, Vol. 51 ›› Issue (11): 182-190.doi: 10.11896/jsjkx.230900022
刘帅, 白雪飞, 高小方
LIU Shuai, BAI Xuefei, GAO Xiaofang
摘要： 针对基于原型网络的小样本学习模型泛化能力不足以及由少量样本得到的类原型不准确等问题,提出一种新的小样本学习方法。首先采用一个由双向卷积块注意力模块和残差块构成的对称网络SCB-Net对图像不同深度的特征进行自适应学习,从而提取到更具代表性的类别特征表示,以有效提高模型的泛化能力;其次提出了一种反欧氏标签传播原型校准算法IELP-PC,利用伪标签策略扩充支持集样本;最后在支持集样本上采用反欧氏距离加权对类原型进行校准,进而提高模型的分类精度。在两个常用数据集mini-ImageNet和tiered-ImageNet上进行了实验,结果验证了所提方法的有效性,与基线模型相比,其在5-way 1-shot上分别提高了6.44%和7.83%,在5-way 5-shot上分别提高了2.68%和2.02%。
中图分类号:
| [1] KRIZHEVSKY A,SUTSKEVER I,HINTON G E.Imagenetclassification with deep convolutional neural networks[J].Communications of the ACM,2017,60(6):84-90. [2] YANG Y,XU Z.Rethinking the value of labels for improving class-imbalanced learning[J].Advances in neural information processing systems,2020,33:19290-19301. [3] GE Y Z,LIU H,WANG Y,et al.Survey on deep learning image recognition in dilemma of small samples[J].Journal of Software,2021,33(1):193-210. [4] ANKOWSKI N,DUCH W,GRA.BCZEWSKI K.Meta-Learning in Computational Intelligence[M].Berlin,Heidelberg:Sprin-ger,2011:97-115. [5] LI F F,FERGUS R,PERONA P.One-shot learning of object categories[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2006,28(4):594-611. [6] WANG Y Q,YAO Q M,KWOK J T,et al.Generalizing from a few examples:A survey on few-shot learning[J].ACM Computing Surveys,2020,53(3):1-34. [7] LI X X,LIU Z Y,WU J J,et al.Total relation network with at-tention for few-shot image classification[J].Chinese Journal of Computers,2023,46(2):371-384. [8] FINN C,ABBEEL P,LEVINE S.Model-agnostic meta-learning for fast adaptation of deep networks[C]//International Confe-renceon Machine Learning.PMLR,2017:1126-1135. [9] HE Y,ZANG C,ZENG P,et al.Convolutional shrinkage neural networks based model-agnostic meta-learning for few-shot learning[J].Neural Processing Letters,2023,55(1):505-518. [10] KIM J,KIM T,KIM S,et al.Edge-labeling graph neural network for few-shot learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2019:11-20. [11] YU T,HE S,SONG Y Z,et al.Hybrid graph neural networks for few-shot learning[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2022:3179-3187. [12] AN S B,GUO Y Q,BAI Y,et al.Survey of few-shot image classification research[J].Journal of Frontiers of Computer Science and Technology,2023,17(3):511-532. [13] ANTONELLI S,AVOLA D,CINQUE L,et al.Few-shot object detection:A survey[J].ACM Computing Surveys,2022,54(11s):1-37. [14] KOCH G,ZEMEL R,SALAKHUTDINOV R.Siamese neuralnetworks for one-shot image recognition[C]//ICMLDeep Learning Workshop.2015. [15] VINYALS O,BLUNDELL C,LILLICRAP T,et al.Matchingnetworks for one shot learning[C]//Proceedings of the 30th International Conference on Neural Information Processing Systems.2016:3637-3645. [16] SNELL J,SWERSKY K,ZEMEL R.Prototypical networks for few-shot learning[C]//Advances in Neural Information Processing Systems.2017:4077-4087. [17] CHEN Y,LIU Z,XU H,et al.Meta-baseline:Exploring simple meta-learning for few-shot learning[C]//Proceedings of the IEEE/CVF international conference on computer vision.2021:9062-9071. [18] XIN S,LIU H.Few-shot Classification based on CBAM andprototype network[C]//2022 4th International Conference on Data-driven Optimization of Complex Systems(DOCS).IEEE,2022:1-6. [19] LIU Y,ZHANG H,YANG Y.Few-Shot Image ClassificationBased on Asymmetric Convolution and Attention Mechanism[C]//2022 4th International Conference on Natural Language Processing(ICNLP).IEEE,2022:217-222. [20] CHEN Z,LIN H,QIANG Z,et al.Image Classification with Frequency Channel Attention under the Few-Shot Condition[C]//2022 IEEE 8th International Conference on Computer and Communications(ICCC).IEEE,2022:1958-1963. [21] SONG J,ZHU Z,LI B,et al.Few-shot Learning based on Multi-Attention and Prototype Correction[C]//2022 8th International Symposium on System Security,Safety,and Reliability(ISSSR).IEEE,2022:83-84. [22] LIU D,BAI L,YU T,et al.Learning a Good Representation for Metric-based Few-shot Classification[C]//2023 15th International Conference on Computer Research and Development(ICCRD).IEEE,2023:187-192. [23] LIU J,SONG L,QIN Y.Prototype rectification for few-shotlearning[C]//Computer Vision-ECCV 2020:16th European Conference.Springer International Publishing,2020:741-756. [24] YANG S,LIU L,XU M.Free lunch for few-shot learning:Distribution calibration[J].arXiv:2101.06395,2021. [25] HUANG Y W,HU Y F,WEI G Q.Prototype-based calibration distribution for few-shot learning[J].Electronic Measurement Technology,2022,45(5):132-139. [26] RUSU A A,RAO D,SYGNOWSKI J,et al.Meta-learning with latent embedding optimization[J].arXiv:1807.05960,2018. [27] LIU Y,SCHIELE B,SUN Q.An ensemble of epoch-wise empi-rical bayes for few-shot learning[C]//Computer Vision-ECCV 2020:16th European Conference.Springer International Publishing,2020:404-421. [28] ORESHKIN B,RODRÍGUEZ LÓPEZ P,LACOSTE A.Tadam:Task dependent adaptive metric for improved few-shot learning[J].Advances in Neural Information Processing Systems,2018,31:719-729. [29] LI H,EIGEN D,DODGE S,et al.Finding task-relevant features for few-shot learning by category traversal[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2019:1-10. [30] LIU Y,LEE J,PARK M,et al.Learning to propagate labels:Transductive propagation network for few-shot learning[J].arXiv:1805.10002,2018. [31] HOU R,CHANG H,MA B,et al.Cross attention network for few-shot classification[J].Advances in Neural Information Processing Systems,2019,32:4003-4014. [32] LV J,ZENG M Y,DONG B S.Prototype rectification few-shot classification model with dual-path co-operation[J].Journal of Frontiers of Computer Science and Technology,2024,18(3):693-706. [33] SELVARAJU R R,COGSWELL M,DAS A,et al.Grad-cam:Visual explanations from deep networks via gradient-based localization[C]//Proceedings of the IEEE International Confe-rence on Computer Vision.2017:618-626. |
|
||
首页