最小差异采样的主动学习图像分类方法
|
摘要
针对委员会成员模型投票不一致性的度量问题,提出了一种基于最小差异采样的主动学习图像分类方法。该方法首先基于标注样本集的重采样结果构建决策委员会,然后利用投票概率较高的2个类别的概率值的差异来度量未标注样本集每个样本的投票不一致性,选择概率差异最小的样本交由人工专家标注,如此迭代更新分类器。将新方法与EQB算法及nEQB算法在多个数据集上进行实验对比,实验结果表明所提方法能够有效提高分类的准确率。还对组成决策委员会的成员模型的数目设置进行了分析和讨论,结果表明在相同的成员模型数目时所提方法比nEQB算法更为有效。
Aiming at the problem of measuring the voting disagreement of committee, a minimal difference sampling method for image classification was proposed. It selects the sample with the minimal difference of two highest class probabilities voted by committee. The experimental results show that this method effectively enhances the classification accuracy compared with EQB and nEQB. Furthermore, the influence of the number of models in the decision-making committee was analyzed and discussed. The experimental results show that the proposed method always outperforms nEQB with the same number of models.
Aiming at the problem of measuring the voting disagreement of committee, a minimal difference sampling method for image classification was proposed. It selects the sample with the minimal difference of two highest class probabilities voted by committee. The experimental results show that this method effectively enhances the classification accuracy compared with EQB and nEQB. Furthermore, the influence of the number of models in the decision-making committee was analyzed and discussed. The experimental results show that the proposed method always outperforms nEQB with the same number of models.
引文
[1]钟桦,杨晓鸣,焦李成.基于多分辨共生矩阵的纹理图像分类[J].计算机研究与发展,2011,48(11):1991-1999.ZHONG H,YANG X M,JIAO L C.Texture classification based on multiresolution co-occurrence matrix[J].Journal of Computer Research and Development,2011,48(11):1991-1999.
[2]CIRESAN D,MEIER U,SCHMIDHUBER J.Multi-column deep neural networks for image classification[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition[C].Rhode Island,USA,2012.3642-3649.
[3]XU Y,ZHU J Y,CHANG E,et al.Multiple clustered instance learning for histopathology cancer image classification,segmentation and clustering[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition(CVPR 2012)[C].Rhode Island,USA,2012.964-971.
[4]VOLPI M,TUIA D,KANEVSKI M.Memory-based cluster sampling for remote sensing image classification[J].IEEE Transactions on Geoscience and Remote Sensing,2012,50(8):3096-3106.
[5]SETTLES B.Active Learning Literature Survey[R].Madison:University of Wisconsin,2010.
[6]LEWIS D D,CATLETT J.Heterogenous uncertainty sampling for supervised learning[A].Proceedings of International Conference on Machine Learning(ICML 1994)[C].New Brunswick,NJ,USA,1994.148-156.
[7]OLSSON F.A Literature Survey of Active Machine Learning in the Context of Natural Language Processing[R].Swedish Institute of Computer Science,2009.
[8]FU Y,ZHU X,LI B.A survey on instance selection for active learning[J].Knowledge and Information Systems,2013,35(2):249-283.
[9]JOSHI A J,PORIKLI F,PAPANIKOLOPOULOS N P.Scalable active learning for multi-class image classification[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(11):2259-2273.
[10]LI X,GUO Y.Adaptive active learning for image classification[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition(CVPR 2013)[C].Portland,Oregon,USA,2013.859-866.
[11]HOI S C H,JIN R,LYU M R.Batch mode active learning with applications to text categorization and image retrieval[J].IEEE Transactions on Knowledge and Data Engineering,2009,21(9):1233-1248.
[12]WANG M,HUA X S.Active learning in multimedia annotation and retrieval:a survey[J].ACM Transactions on Intelligent Systems and Technology,2011,2(2):1899412-1899414.
[13]梁爽,孙正兴.面向草图检索的小样本增量有偏学习算法[J].软件学报,2009,20(5):1301-1312.LIANG S,SUN Z X.Small sample incremental biased learning algorithm for sketch retrieval[J].Journal of Software,2009,20(5):1301-1312.
[14]LEWIS D D,GALE W A.A sequential algorithm for training text classifiers[A].Proceedings of the 17th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval[C].Dublin,Ireland,1994.3-12.
[15]SEUNG H S,OPPER M,SOMPOLINSKY H.Query by committee[A].Proceedings of the Fifth Annual Workshop on Computational Learning Theory[C].Pittsburgh,PA,USA,1992.287-294.
[16]吴伟宁,刘扬,郭茂祖等.基于采样策略的主动学习算法研究进展[J].计算机研究与发展,2012,49(6):1162-1173.WU W N,LIU Y,GUO M Z,et al.Advances in active learning algorithms based on sampling strategy[J].Journal of Computer Research and Development,2012,49(6):1162-1173.
[17]ABE N,MAMITSUKA H.Query learning strategies using boosting and bagging[A].Proceedings of the Fifteenth International Conference(ICML'98)[C].Madison,Wisconsin,USA,1998.1-9.
[18]TUIA D,RATLE F,PACIFICI F,et al.Active learning methods for remote sensing image classification[J].IEEE Transactions on Geoscience and Remote Sensing,2009,47(7):2218-2232.
[19]COPA L,TUIA D,VOLPI M,et al.Unbiased query-by-bagging active learning for VHR image classification[A].Proceedings of SPIE Remote Sensing[C].Toulouse,France,2010.783001-783008.
[20]BREIMAN L.Bagging predictors[J].Machine learning,1996,24(2):123-140.
[21]ASUNCION A,NEWMAN D J.UCI machine learning repository.[EB/OL].http://www.ics.uci.edu/mlearn/MLRepository.html,2007.
[22]COLLOBERT R,BENGIO S,MARIETHOZ J.Torch:A Modular Machine Learning Software Library[R].Technical Report,2002.
[2]CIRESAN D,MEIER U,SCHMIDHUBER J.Multi-column deep neural networks for image classification[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition[C].Rhode Island,USA,2012.3642-3649.
[3]XU Y,ZHU J Y,CHANG E,et al.Multiple clustered instance learning for histopathology cancer image classification,segmentation and clustering[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition(CVPR 2012)[C].Rhode Island,USA,2012.964-971.
[4]VOLPI M,TUIA D,KANEVSKI M.Memory-based cluster sampling for remote sensing image classification[J].IEEE Transactions on Geoscience and Remote Sensing,2012,50(8):3096-3106.
[5]SETTLES B.Active Learning Literature Survey[R].Madison:University of Wisconsin,2010.
[6]LEWIS D D,CATLETT J.Heterogenous uncertainty sampling for supervised learning[A].Proceedings of International Conference on Machine Learning(ICML 1994)[C].New Brunswick,NJ,USA,1994.148-156.
[7]OLSSON F.A Literature Survey of Active Machine Learning in the Context of Natural Language Processing[R].Swedish Institute of Computer Science,2009.
[8]FU Y,ZHU X,LI B.A survey on instance selection for active learning[J].Knowledge and Information Systems,2013,35(2):249-283.
[9]JOSHI A J,PORIKLI F,PAPANIKOLOPOULOS N P.Scalable active learning for multi-class image classification[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(11):2259-2273.
[10]LI X,GUO Y.Adaptive active learning for image classification[A].Proceedings of IEEE Conference on Computer Vision and Pattern Recognition(CVPR 2013)[C].Portland,Oregon,USA,2013.859-866.
[11]HOI S C H,JIN R,LYU M R.Batch mode active learning with applications to text categorization and image retrieval[J].IEEE Transactions on Knowledge and Data Engineering,2009,21(9):1233-1248.
[12]WANG M,HUA X S.Active learning in multimedia annotation and retrieval:a survey[J].ACM Transactions on Intelligent Systems and Technology,2011,2(2):1899412-1899414.
[13]梁爽,孙正兴.面向草图检索的小样本增量有偏学习算法[J].软件学报,2009,20(5):1301-1312.LIANG S,SUN Z X.Small sample incremental biased learning algorithm for sketch retrieval[J].Journal of Software,2009,20(5):1301-1312.
[14]LEWIS D D,GALE W A.A sequential algorithm for training text classifiers[A].Proceedings of the 17th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval[C].Dublin,Ireland,1994.3-12.
[15]SEUNG H S,OPPER M,SOMPOLINSKY H.Query by committee[A].Proceedings of the Fifth Annual Workshop on Computational Learning Theory[C].Pittsburgh,PA,USA,1992.287-294.
[16]吴伟宁,刘扬,郭茂祖等.基于采样策略的主动学习算法研究进展[J].计算机研究与发展,2012,49(6):1162-1173.WU W N,LIU Y,GUO M Z,et al.Advances in active learning algorithms based on sampling strategy[J].Journal of Computer Research and Development,2012,49(6):1162-1173.
[17]ABE N,MAMITSUKA H.Query learning strategies using boosting and bagging[A].Proceedings of the Fifteenth International Conference(ICML'98)[C].Madison,Wisconsin,USA,1998.1-9.
[18]TUIA D,RATLE F,PACIFICI F,et al.Active learning methods for remote sensing image classification[J].IEEE Transactions on Geoscience and Remote Sensing,2009,47(7):2218-2232.
[19]COPA L,TUIA D,VOLPI M,et al.Unbiased query-by-bagging active learning for VHR image classification[A].Proceedings of SPIE Remote Sensing[C].Toulouse,France,2010.783001-783008.
[20]BREIMAN L.Bagging predictors[J].Machine learning,1996,24(2):123-140.
[21]ASUNCION A,NEWMAN D J.UCI machine learning repository.[EB/OL].http://www.ics.uci.edu/mlearn/MLRepository.html,2007.
[22]COLLOBERT R,BENGIO S,MARIETHOZ J.Torch:A Modular Machine Learning Software Library[R].Technical Report,2002.