基于空间稀疏编码的MIL算法及刑侦图像分类
|
摘要
针对刑侦图像分类问题,提出一种基于空间稀疏编码(SSC)的多示例学习(MIL)算法。首先,利用稠密尺度不变特征转换(SIFT)原理设计一种带有示例位置信息的多示例建模方案,将刑侦图像分类问题转化为MIL问题;然后,基于多样性密度(DD)函数及稀疏编码(SC)理论,设计了一种针对MIL的字典构造方法及空间稀疏编码方案,用于计算多示例包的元数据(metadata);最后,结合大尺度线性支持向量机方法,提出了一种SSC-MIL的MIL新算法。14类真实刑侦图像的对比实验表明,该算法是有效的,且分类精度高于其他方法。
Focusing on the classification problem of the criminal investigation, a multi-instance learning(MIL) algorithm based on spatial sparse coding(SSC) is proposed. By using the dense scale invariant feature transform(SIFT) principle, a multi-instance modeling scheme with instance position information is constructed to transform the problem of criminal investigation image classification into a multi-instance learning(MIL) problem. Based on the diversity density(DD) function and the sparse coding theory, a new dictionary construct method and spatially sparse coding(SSC) technique are designed for MIL, to extract the metadata for each multi-instance bag. At last, a new MIL algorithm called SSC-MIL is proposed by combining the large-scale linear support vector machine method. Experimental results on the 14 cases of real criminal investigation image show that the proposed method is effective, and the classification accuracy is higher than other methods.
Focusing on the classification problem of the criminal investigation, a multi-instance learning(MIL) algorithm based on spatial sparse coding(SSC) is proposed. By using the dense scale invariant feature transform(SIFT) principle, a multi-instance modeling scheme with instance position information is constructed to transform the problem of criminal investigation image classification into a multi-instance learning(MIL) problem. Based on the diversity density(DD) function and the sparse coding theory, a new dictionary construct method and spatially sparse coding(SSC) technique are designed for MIL, to extract the metadata for each multi-instance bag. At last, a new MIL algorithm called SSC-MIL is proposed by combining the large-scale linear support vector machine method. Experimental results on the 14 cases of real criminal investigation image show that the proposed method is effective, and the classification accuracy is higher than other methods.
引文
[1]曹健,魏星,李海生,等.基于局部调整的图像分类方法[J].电子科技大学学报,2017,46(1):69-74.CAO Jian,WEI Xing,LI Hai-sheng,et al.Image classification methods based on local feature[J].Journal of University of Electronic Science and Technology of China,2017,46(1):69-74.
[2]韩德明.信息化背景下侦查权范式的要素系谱[J].中国人民公安大学学报(社会科学版),2016,182(4):66-72.HAN De-ming.Essential genealogy of the right of investigation in the context of information[J].Journal of People's Public Security University of China(Social Science Edition),2016,182(4):66-72.
[3]GWO C Y,WEI C H.Shoeprint retrieval:Core point alignment for pattern comparison[J].Science&Justice,2016,56(5):341-350.
[4]MEDINA M A,MORENO A M,BALLESTER M A F,et al.Latent fingerprint identification using deformable minutiae clustering[J].Neurocomputing,2016,175(1):851-865.
[5]LIU Y,YAN H,LIM K P.Study on rotation-invariant texture feature extraction for tire pattern retrieval[J].Multidimensional Systems&Signal Processing,2015,21(2):1-14.
[6]YANG J,YU K,GONG Y,et al.Linear spatial pyramid matching using sparse coding for image classification[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.[S.l.]:IEEE,2009:1794-1801.
[7]李大湘,赵小强,李娜.图像语义分析的MIL算法综[J].控制与决策,2013,28(4):481-488.LI Da-xiang,ZHAO Xiao-qiang,LI Na.A survey of MILalgorithm for image semantic analysis[J].Control and Decision Making,2013,28(4):481-488.
[8]MARON O,RATAN A L.Multiple-instance learning for natural scene classification[C]//Proceedings of the 15th International Conference on Machine Learning.Sab Frabcuscim,CA,USA:Morgan Kaufmann Publishers Inc,1998:341-349.
[9]ANDREWS S,TSOCHANTARIDIS I,HOFMANN T.Support vector machines for multiple-instance learning[C]//Proceedings of the 15th Neural Information Processing Systems.Cambridge,MA,USA:MIT,2003:561-568.
[10]ZHOU Z H,SUN Y Y,LI Y F.Multi-Instance learning by treating instances as non-I.I.D.samples[C]//Proceedings of the 26th International Conference on Machine Learning.Montreal,Canada:[s.n.],2009:1249-1256.
[11]CHEN Y,WANG J Z.Image categorization by learning and reasoning with regions[J].Journal of Machine Learning Research,2004,5(8):913-939.
[12]CHEN Y,BI J,WANG J Z.MILES:Multiple-Instance learning via embedded instance selection[J].IEEETransactions on Pattern Analysis and Machine Intelligence,2006,28(12):1931-1947.
[13]WEI X S,WU J,ZHOU Z H.Scalable algorithms for multi-instance learning[J].IEEE Transactions on Neural Networks and Learning Systems,2017,28(4):975-987.
[14]LI D X,PENG J Y,LI Z,et al.LSA based multi-instance learning algorithm for image retrieval[J].Signal Processing,2011,91(8):1993-2000.
[15]AMORES J.Multiple instance classification:Review,taxonomy and comparative study[J].Artificial Intelligence,2013,201(4):81-105.
[16]SHRIVASTAVA A,PATEL V M,PILLAI J K,et al.Generalized dictionaries for multiple instance learning[J].International Journal of Computer Vision,2015,114(2-3):288-305.
[17]李宗民,蒋迪,刘玉杰,等.结合空间上下文的局部约束线性特征编码[J].计算机辅助设计与图形学学报,2017,29(2):254-261.LI Zong-min,JIANG Di,LIU Yu-jie,et al.Local constrained linear feature coding based on spatial context[J].Journal of Computer Aided Design&Computer Graphics,2017,29(2):254-261.
[18]FAN R E,CHANG K W,HSIEH C J,et al.LIBLINEAR:a library for large linear classification[EB/OL].[2017-02-21].http://www.csie.ntu.edu.tw/~cjlin/liblinear/.
[2]韩德明.信息化背景下侦查权范式的要素系谱[J].中国人民公安大学学报(社会科学版),2016,182(4):66-72.HAN De-ming.Essential genealogy of the right of investigation in the context of information[J].Journal of People's Public Security University of China(Social Science Edition),2016,182(4):66-72.
[3]GWO C Y,WEI C H.Shoeprint retrieval:Core point alignment for pattern comparison[J].Science&Justice,2016,56(5):341-350.
[4]MEDINA M A,MORENO A M,BALLESTER M A F,et al.Latent fingerprint identification using deformable minutiae clustering[J].Neurocomputing,2016,175(1):851-865.
[5]LIU Y,YAN H,LIM K P.Study on rotation-invariant texture feature extraction for tire pattern retrieval[J].Multidimensional Systems&Signal Processing,2015,21(2):1-14.
[6]YANG J,YU K,GONG Y,et al.Linear spatial pyramid matching using sparse coding for image classification[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.[S.l.]:IEEE,2009:1794-1801.
[7]李大湘,赵小强,李娜.图像语义分析的MIL算法综[J].控制与决策,2013,28(4):481-488.LI Da-xiang,ZHAO Xiao-qiang,LI Na.A survey of MILalgorithm for image semantic analysis[J].Control and Decision Making,2013,28(4):481-488.
[8]MARON O,RATAN A L.Multiple-instance learning for natural scene classification[C]//Proceedings of the 15th International Conference on Machine Learning.Sab Frabcuscim,CA,USA:Morgan Kaufmann Publishers Inc,1998:341-349.
[9]ANDREWS S,TSOCHANTARIDIS I,HOFMANN T.Support vector machines for multiple-instance learning[C]//Proceedings of the 15th Neural Information Processing Systems.Cambridge,MA,USA:MIT,2003:561-568.
[10]ZHOU Z H,SUN Y Y,LI Y F.Multi-Instance learning by treating instances as non-I.I.D.samples[C]//Proceedings of the 26th International Conference on Machine Learning.Montreal,Canada:[s.n.],2009:1249-1256.
[11]CHEN Y,WANG J Z.Image categorization by learning and reasoning with regions[J].Journal of Machine Learning Research,2004,5(8):913-939.
[12]CHEN Y,BI J,WANG J Z.MILES:Multiple-Instance learning via embedded instance selection[J].IEEETransactions on Pattern Analysis and Machine Intelligence,2006,28(12):1931-1947.
[13]WEI X S,WU J,ZHOU Z H.Scalable algorithms for multi-instance learning[J].IEEE Transactions on Neural Networks and Learning Systems,2017,28(4):975-987.
[14]LI D X,PENG J Y,LI Z,et al.LSA based multi-instance learning algorithm for image retrieval[J].Signal Processing,2011,91(8):1993-2000.
[15]AMORES J.Multiple instance classification:Review,taxonomy and comparative study[J].Artificial Intelligence,2013,201(4):81-105.
[16]SHRIVASTAVA A,PATEL V M,PILLAI J K,et al.Generalized dictionaries for multiple instance learning[J].International Journal of Computer Vision,2015,114(2-3):288-305.
[17]李宗民,蒋迪,刘玉杰,等.结合空间上下文的局部约束线性特征编码[J].计算机辅助设计与图形学学报,2017,29(2):254-261.LI Zong-min,JIANG Di,LIU Yu-jie,et al.Local constrained linear feature coding based on spatial context[J].Journal of Computer Aided Design&Computer Graphics,2017,29(2):254-261.
[18]FAN R E,CHANG K W,HSIEH C J,et al.LIBLINEAR:a library for large linear classification[EB/OL].[2017-02-21].http://www.csie.ntu.edu.tw/~cjlin/liblinear/.