面向开放集图像分类的模糊域自适应方法

doi:10.3778/j.issn.1673-9418.1912036

摘要/Abstract

摘要：

目前大多的域自适应算法在源域与目标域具有相同类别的场景下，利用标签丰富的源域信息对标签稀少且分布相似的目标域数据进行迁移学习，取得了很多成果。然而，由于现实场景的复杂性和开放性，源域和目标域在类别空间上不尽相同，往往会各自包含一些类别未知且超出现有类别设定的样本。对于这样具有挑战性的开放集场景，传统的域自适应算法将无能为力。为了有效解决上述问题，提出一种面向开放集的模糊域自适应算法。该算法引用了不确定性的模糊化，计算目标域样本的模糊隶属度来学习源域特征到目标域特征空间的线性映射，通过迭代逐步将源域与目标域转化在同一特征空间下。通过对无监督和半监督的图像迁移任务的大量实验，验证了该算法对于开放集场景下图像分类的有效性。

关键词: 开放集合, 域自适应, 模糊隶属度, 图像分类

Abstract:

Key words: 开放集合, 域自适应, 模糊隶属度, 图像分类

刘晓龙, 王士同. 面向开放集图像分类的模糊域自适应方法[J]. 计算机科学与探索, 2021, 15(3): 515-523.

LIU Xiaolong, WANG Shitong. Fuzzy Domain Adaptation Algorithm for Open Set Image Classification[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(3): 515-523.

参考文献

[1] SCHEIRER W J, DE REZENDE ROCHA A, SAPKOTA A, et al. Toward open set recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35(7): 1757-1772.
[2] TZENG E, HOFFMAN J, SAENKO K, et al. Adversarial discriminative domain adaptation[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 2962-2971.
[3] MOTIIAN S, PICCIRILLI M, ADJEROH D A, et al. Uni-fied deep supervised domain adaptation and generalization[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 5715-5725.
[4] HSU T M, CHEN W Y, HOU C A, et al. Unsupervised domain adaptation with imbalanced cross-domain data[C]//Proceedings of the 2015 IEEE International Conference on Computer Vision, Santiago, Dec 7-13, 2015. Washington: IEEE Computer Society, 2015: 4121-4129.
[5] LIU J W, SUN Z K, LUO X L. Review and research develop-ment on domain adaptation learning[J]. Acta Automatica Sinica, 2014, 40(8): 1576-1600.
刘建伟, 孙正康, 罗雄麟. 域自适应学习研究进展[J]. 自动化学报, 2014, 40(8): 1576-1600.
[6] GANIN Y, LEMPITSKY V. Unsupervised domain adaptation by backpropagation[J]. arXiv:1409.7495, 2014.
[7] PAN S J, TSANG I W, KWOK J T, et al. Domain adapta-tion via transfer component analysis[J]. IEEE Transactions on Neural Networks, 2010, 22(2): 199-210.
[8] WANG J D, CHEN Y Q, HAO S J, et al. Balanced distribu-tion adaptation for transfer learning[C]//Proceedings of the 2017 IEEE International Conference on Data Mining, New Orleans, Nov 18-21, 2017. Washington: IEEE Computer Society, 2017: 1129-1134.
[9] GONG B Q, SHI Y, SHA F, et al. Geodesic flow kernel for unsupervised domain adaptation[C]//Proceedings of the 2012 IEEE Conference on Computer Vision and Pattern Reco-gnition, Providence, Jun 16-21, 2012. Washington: IEEE Com-puter Society, 2012: 2066-2073.
[10] GOPALAN R, LI R N, CHELLAPPA R. Domain adapta-tion for object recognition: an unsupervised approach[C]//Proceedings of the 2011 IEEE International Conference on Computer Vision, Barcelona, Nov 6-13, 2011. Washington:IEEE Computer Society, 2011: 999-1006.
[11] SUN B C, FENG J S, SAENKO K. Return of frustratingly easy domain adaptation[C]//Proceedings of the 30th AAAI Conference on Artificial Intelligence, Phoenix, Feb 12-17, 2016. Menlo Park: AAAI, 2016: 2058-2065.
[12] SUN B C, SAENKO K. Deep coral: correlation alignment for deep domain adaptation[C]//LNCS 9915: Proceedings of the European Conference on Computer Vision, Amsterdam, Oct 8-10 and 15-16, 2016. Berlin, Heidelberg: Springer, 2016: 443-450.
[13] WANG J, WANG S T. Double indices FCM algorithm based on hybrid distance metric learning[J]. Journal of Soft-ware, 2010, 21(8): 1878-1888.
王骏, 王士同. 基于混合距离学习的双指数模糊C均值算法[J]. 软件学报, 2010, 21(8): 1878-1888.
[14] JIA H N, WANG S T. Reinforced rule-based fuzzy models for noisy data and its implementation[J]. Journal of Frontiers of Computer Science and Technology, 2018, 12(11): 120-131.
贾海宁, 王士同. 面向噪声数据的强化模糊规则模型及实现[J]. 计算机科学与探索, 2018, 12(11): 120-131.
[15] KESHAVARZ A, WANG Y, BOYD S P. Imputing a convex objective function[C]//Proceedings of the 2011 IEEE Inter-national Symposium on Intelligent Control, Denver, Sep 28-30, 2011. Piscataway: IEEE, 2011: 613-619.
[16] FERNANDO B, HABRARD A, SEBBAN M, et al. Unsu-pervised visual domain adaptation using subspace alignment[C]//Proceedings of the 2013 IEEE International Conference on Computer Vision, Sydney, Dec 1-8, 2013. Washington:IEEE Computer Society, 2013: 2960-2967.
[17] CHANG C C, LIN C J. LIBSVM: a library for support vector machines[J]. ACM Transactions on Intelligent Systems and Technology, 2011, 2(3): 27.
[18] SAENKO K, KULIS B, FRITZ M, et al. Adapting visual category models to new domains[C]//LNCS 6314: Proceed-ings of the European Conference on Computer Vision, Hera-klion, Sep 5-11, 2010. Berlin, Heidelberg: Springer, 2010: 213-226.
[19] PANAREDA BUSTO P, GALL J. Open set domain adap-tation[C]//Proceedings of the 2017 IEEE International Con-ference on Computer Vision, Venice, Oct 22-29, 2017. Wash-ington: IEEE Computer Society, 2017: 754-763.
[20] SAITO K, YAMAMOTO S, USHIKU Y, et al. Open set domain adaptation by backpropagation[C]//LNCS 11209:Proceedings of the European Conference on Computer Vision, Munich, Sep 8-14, 2018. Berlin, Heidelberg: Springer, 2018: 156-171.
[21] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[C]//Proceedings of the 26th Annual Conference on Neural Infor-mation Processing Systems, Lake Tahoe, Dec 3-6, 2012. Red Hook: Curran Associates, 2012: 1097-1105.