多尺度表达和正则化方法在图像识别中的研究与应用
|
摘要
随着移动设备的不断普及和互联网的持续发展，以图像为载体的信息交流方式越来越普遍。如何能够自动理解拍摄图像传达的语义信息是人们实际且迫切的需求。图像识别作为解决这类问题不可或缺的手段，主要关心如何对图像中的物体及对其所处的场景进行分析判断。通常，物体识别包括个体识别、次级类别识别以及类别间的识别，而场景识别(又称为场景分类)则是根据图像内容给出其代表的语义标签(如山脉、海岸线等)，可为物体识别等任务提供有效的上下文信息。为解决这两类模式识别问题，本文从对图像识别系统关键模块的现状分析出发，重点研究了特征设计中的多尺度信息和方法论中的正则化手段在图像光照预处理、人脸性别识别、物体识别和场景识别中的应用。
首先，针对光照条件对图像识别的影响，本文考虑解决光源变化带来的图像颜色漂移问题，以达到色彩恒常目的。目前，不存在任何一种通用的单个色彩恒常算法能适用于不同纹理特性的图像。为此，提出一种基于纹理金字塔特征匹配和正则化局部回归(TPM RLR)的色彩恒常算法来融合单个算法的光照估计结果。TPM RLR算法结合多尺度表达构建纹理金字塔，并利用威布尔分布参数提取图像的纹理特征。然后，根据自定义的图像相似性测度查找到与待校正图像纹理最相似的参考图像集。在融合阶段，综合基于先验知识的方法与基于数据驱动方法的优点，在lαβ对立色彩空间内采用正则化局部回归合并单个算法。在两个自然图像数据集上的实验结果表明，TPM RLR融合算法能显著提高光照估计的准确度，其角度误差中值比单个算法至少降低了29%，同时其校正效果在主观和客观评价上均优于现有的融合算法。
其次，以人脸这种特殊物体为研究对象，讨论了基于人脸的性别识别问题。首先，提出一种基于多尺度学习模式(MSLP)的人脸特征，该特征通过不同方法PCA/ICA/K-means学习得到多尺度卷积模板，然后对卷积后的图像按照响应强度顺序进行编码，从而形成一个紧凑且具有较高鉴别力的直方图特征来描述人脸。鉴于线性表达分类器(LinearRC)在人脸识别中的成功应用，同时又考虑到性别识别和人脸识别在数据分布上的差异，系统分析了LinearRC在性别识别上的应用。另外，从原型生成的角度出发，提出一种基于偏最小二乘的线性表达分类器(LRC PLS)及其群组分类器。在人脸性别识别上的大量实验证明，MSLP特征明显优于人工设计特征，LRC PLS较其他线性表达分类器性能稳定、耗时较短，且其群组分类器可进一步提高识别精度。
鉴于原型概念在性别识别中的成功应用，从字典学习角度出发，进而提出一种基于多尺度查询图像扩展的协同表达分类器(QCRC CP)，以解决较复杂环境下的物体识别问题(即处理多姿态的、多视角的、更一般化等情况下的物体识别)。QCRC CP先将单张查询图像通过放缩扩展成一个查询集，并利用典型相关分析生成与查询集最相关的一组类原型作为每一类的字典原子。这种字典构造方法结合以往原型选择方法与原型生成方法的优点，利用数据局部性剔除掉噪声，生成与查询样本相关且最能代表该类别的原型集。最后，在新构造的字典上，采用基于最小规范化残差的多变量协同表达分类器判定查询图像的类别。在多姿态人脸识别、树叶类别识别、字符识别和通用物体识别上的实验表明，QCRC CP分类器获得了令人满意的结果，且其字典构造方法优于基于原型选择和原型生成的字典学习算法。例如，QCRC CP在字符识别上的识别率比其他线性表达分类器提高了10%以上。
最后，在场景识别中，考虑到合适的距离度量能有效反映高维空间中样本间的语义距离，提出一种基于正则化线性判别分析(RLDA)的全参数马氏距离度量学习算法。当训练样本数目少且特征维数高时，距离度量学习算法复杂度过高，且容易出现过拟合问题。为此，提出一种新的算法框架，即将度量学习的参数矩阵分解为一个射影矩阵与一个非负对角阵的乘积，从而可显著降低要估计的参数个数。首先，采用不同参数调制的RLDA生成一个射影向量集合。然后基于边信息，在构建的训练数据集上通过l2范数正则化的非负最小二乘对射影向量进行选择和加权。该训练集是由射影后的相似样本对子集与相异样本对子集中样本对的差向量平方组成。另外，为了保持训练库中两个子集样本数目的相对平衡，提出一种基于K近邻的简单却有效的策略。在两个场景图像集上的实验表明，所提方法较其他度量学习方法能够获得更好的场景识别效果，且运行速度是传统度量学习方法的几倍、甚至几十倍。
With the nonstop popularization of mobile devices and the continuous advancesof the Internet, the information exchanges with images as the carrier become moreprevalent. How to automatically understand what a taken photo conveys at the seman-tic level is a both pragmatic and imperative demand. Image recognition, which is ofessence regarding the recognition of objects in the image and the scene the objects re-side in, serves as an indispensable tool. Generally, object recognition ranges from therecognition of the individual, to subordinate categories and to basic-level categories.Scene recognition (a.k.a, scene categorization) is to identify the semantic label (e.g.,mountain, coast, etc.) of an image via its content and thus provides the valid contextfor tasks like object recognition. To solve these two kinds of pattern recognition prob-lems, with an analysis on the status of key modules of an image recognition system,the dissertation puts an emphasis on the multi-scale information in devising a featureand regularization in constructing an effective method and conducts a broad researchof their applications in illumination preprocessing for image recognition, facial genderrecognition, object recognition, and scene recognition.
With a view to the influence of illumination on image recognition, the first prob-lem under consideration is to solve the color shift caused by illuminant change forthe purpose of color constancy. In general, single color constancy algorithms cannotconstantly gain perfect results on the images with various texture for their specific as-sumptions. A fusion method based on texture pyramid matching and regularized localregression (TPM RLR) is proposed accordingly. We first construct a texture pyra-mid based on multi-scale representation and use the Weibull distribution parametersto extract texture features of images. Then we define a new image similarity measureto retrieve similar images of the test image. Finally, considering the integration of data-driven methods and prior-knowledge-based methods, we apply a regularized lo-cal regression to fuse the results of single algorithms in an opponent color space, i.e.,lαβ. The experimental results on two natural image datasets indicate TPM RLR canstrikingly improve the illumination estimation with at least29%decrease w.r.t medi-an angular error compared to the best-performing single algorithm, and its correctedeffect is superior to those of other combination methods under either subjective andobjective criteria.
Focusing on a specific object, i.e., face, we consider the facial gender recogni-tion as the next topic. First, we propose a multi-scale learned pattern (MSLP) basedface descriptor. The proposed feature learns multi-scale convolution templates via oneof algorithms like ICA, PCA or K-means, then encodes the images according to theresponse order of these templates, and finally forms a more compact and discrimi-native histogram feature for facial representation. Having witnessed the success oflinear representation based classifiers (LinearRCs) in face recognition and consideredthe different characteristics of data distribution between face recognition and genderrecognition, we systematically analyze the application of LinearRCs on facial gen-der recognition. Stemming from the idea of prototype generation, a LinearRC usingpartial least squares (i.e., LRC PLS) and its group based variant are proposed. The ex-tensive experiments on facial gender recognition show that the features using MSLPare better than previous hand-engineered features, that LRC PLS is more stable thanother LinearRCs while having relatively less prediction time, and that its group variantcan further boost the performance.
What’s more, in light of the success of prototype concept in gender recognition,the dissertation presents a LinearRC named multi-scale query-expanded collabora-tive representation based classifier with class-specific prototypes (QCRC CP) fromthe viewpoint of dictionary learning in order to deal with more complex object recog-nition (i.e., to deal with objects with multiple pose or view, more generic objects, etc.).We firstly expand a single query image into a query set by scaling, and then constructa query-dependent dictionary via canonical correlation analysis (CCA) between thequery set and the samples from a specific class. By integrating the prototype selectionbased methods and prototype generation based methods, the proposed method of dic-tionary learning generates query-related class-specific prototypes in a class-wise way, in which the implicit query-dependent data locality discards the outliers. In the end,multivariate collaborative representation based classification (CRC) using newly con-structed dictionary is employed to determine the identity according to the rule of mini-mum normalized residual (MNR). The experiments cover multi-pose face recognition,leaf species recognition, character recognition in natural scene and generic multi-viewobject recognition. The results demonstrate that QCRC CP can gain pleasing resultsand the proposed approach of dictionary learning is superior to previous methods viaprototype selection or generation. For instance, in contrast to other LinearRCs, QCR-C CP improves the recognition accuracy with an increase of over10%on characterrecognition.
Last but not least, considering a proper distance metric in scene recognition caneffectively reflect semantic distance between samples in high-dimensional space, wepropose a metric learning algorithm based on regularized linear discriminant anal-ysis (RLDA) to learn a full parameter matrix of Mahalanobis distance metric. It’sacknowledged that high-dimensional features with relatively limited samples mightintroduce huge computational cost and overfitting in conventional metric learning. Anovel framework is thus presented to resolve these problems. In our framework, theestimation of the parameter matrix is decomposed into that of a projection pool ma-trix and a non-negative diagonal selection matrix, which can significantly reduce thenumber of estimated parameters to a small scale. We first construct the projection vec-tors by tuning regularized LDA with different parameters. Based on side information,nonnegative l2-norm regularized least squares is then used to select and weigh the pro-jection vectors on the constructed training dataset. The training dataset is comprisedof the pairwise squared differences of the projected samples in the dissimilarity andsimilarity pair subsets. To maintain the balance of two subsets, a simple but effective s-trategy based on K nearest neighbors (KNN) is provided. It turns out that the proposedalgorithm can achieve higher recognition rates on two datasets for scene recognitionwhile retaining such high efficiency that obtains several-fold, and even dozens-foldacceleration compared with conventional metric learning methods.
首先，针对光照条件对图像识别的影响，本文考虑解决光源变化带来的图像颜色漂移问题，以达到色彩恒常目的。目前，不存在任何一种通用的单个色彩恒常算法能适用于不同纹理特性的图像。为此，提出一种基于纹理金字塔特征匹配和正则化局部回归(TPM RLR)的色彩恒常算法来融合单个算法的光照估计结果。TPM RLR算法结合多尺度表达构建纹理金字塔，并利用威布尔分布参数提取图像的纹理特征。然后，根据自定义的图像相似性测度查找到与待校正图像纹理最相似的参考图像集。在融合阶段，综合基于先验知识的方法与基于数据驱动方法的优点，在lαβ对立色彩空间内采用正则化局部回归合并单个算法。在两个自然图像数据集上的实验结果表明，TPM RLR融合算法能显著提高光照估计的准确度，其角度误差中值比单个算法至少降低了29%，同时其校正效果在主观和客观评价上均优于现有的融合算法。
其次，以人脸这种特殊物体为研究对象，讨论了基于人脸的性别识别问题。首先，提出一种基于多尺度学习模式(MSLP)的人脸特征，该特征通过不同方法PCA/ICA/K-means学习得到多尺度卷积模板，然后对卷积后的图像按照响应强度顺序进行编码，从而形成一个紧凑且具有较高鉴别力的直方图特征来描述人脸。鉴于线性表达分类器(LinearRC)在人脸识别中的成功应用，同时又考虑到性别识别和人脸识别在数据分布上的差异，系统分析了LinearRC在性别识别上的应用。另外，从原型生成的角度出发，提出一种基于偏最小二乘的线性表达分类器(LRC PLS)及其群组分类器。在人脸性别识别上的大量实验证明，MSLP特征明显优于人工设计特征，LRC PLS较其他线性表达分类器性能稳定、耗时较短，且其群组分类器可进一步提高识别精度。
鉴于原型概念在性别识别中的成功应用，从字典学习角度出发，进而提出一种基于多尺度查询图像扩展的协同表达分类器(QCRC CP)，以解决较复杂环境下的物体识别问题(即处理多姿态的、多视角的、更一般化等情况下的物体识别)。QCRC CP先将单张查询图像通过放缩扩展成一个查询集，并利用典型相关分析生成与查询集最相关的一组类原型作为每一类的字典原子。这种字典构造方法结合以往原型选择方法与原型生成方法的优点，利用数据局部性剔除掉噪声，生成与查询样本相关且最能代表该类别的原型集。最后，在新构造的字典上，采用基于最小规范化残差的多变量协同表达分类器判定查询图像的类别。在多姿态人脸识别、树叶类别识别、字符识别和通用物体识别上的实验表明，QCRC CP分类器获得了令人满意的结果，且其字典构造方法优于基于原型选择和原型生成的字典学习算法。例如，QCRC CP在字符识别上的识别率比其他线性表达分类器提高了10%以上。
最后，在场景识别中，考虑到合适的距离度量能有效反映高维空间中样本间的语义距离，提出一种基于正则化线性判别分析(RLDA)的全参数马氏距离度量学习算法。当训练样本数目少且特征维数高时，距离度量学习算法复杂度过高，且容易出现过拟合问题。为此，提出一种新的算法框架，即将度量学习的参数矩阵分解为一个射影矩阵与一个非负对角阵的乘积，从而可显著降低要估计的参数个数。首先，采用不同参数调制的RLDA生成一个射影向量集合。然后基于边信息，在构建的训练数据集上通过l2范数正则化的非负最小二乘对射影向量进行选择和加权。该训练集是由射影后的相似样本对子集与相异样本对子集中样本对的差向量平方组成。另外，为了保持训练库中两个子集样本数目的相对平衡，提出一种基于K近邻的简单却有效的策略。在两个场景图像集上的实验表明，所提方法较其他度量学习方法能够获得更好的场景识别效果，且运行速度是传统度量学习方法的几倍、甚至几十倍。
With the nonstop popularization of mobile devices and the continuous advancesof the Internet, the information exchanges with images as the carrier become moreprevalent. How to automatically understand what a taken photo conveys at the seman-tic level is a both pragmatic and imperative demand. Image recognition, which is ofessence regarding the recognition of objects in the image and the scene the objects re-side in, serves as an indispensable tool. Generally, object recognition ranges from therecognition of the individual, to subordinate categories and to basic-level categories.Scene recognition (a.k.a, scene categorization) is to identify the semantic label (e.g.,mountain, coast, etc.) of an image via its content and thus provides the valid contextfor tasks like object recognition. To solve these two kinds of pattern recognition prob-lems, with an analysis on the status of key modules of an image recognition system,the dissertation puts an emphasis on the multi-scale information in devising a featureand regularization in constructing an effective method and conducts a broad researchof their applications in illumination preprocessing for image recognition, facial genderrecognition, object recognition, and scene recognition.
With a view to the influence of illumination on image recognition, the first prob-lem under consideration is to solve the color shift caused by illuminant change forthe purpose of color constancy. In general, single color constancy algorithms cannotconstantly gain perfect results on the images with various texture for their specific as-sumptions. A fusion method based on texture pyramid matching and regularized localregression (TPM RLR) is proposed accordingly. We first construct a texture pyra-mid based on multi-scale representation and use the Weibull distribution parametersto extract texture features of images. Then we define a new image similarity measureto retrieve similar images of the test image. Finally, considering the integration of data-driven methods and prior-knowledge-based methods, we apply a regularized lo-cal regression to fuse the results of single algorithms in an opponent color space, i.e.,lαβ. The experimental results on two natural image datasets indicate TPM RLR canstrikingly improve the illumination estimation with at least29%decrease w.r.t medi-an angular error compared to the best-performing single algorithm, and its correctedeffect is superior to those of other combination methods under either subjective andobjective criteria.
Focusing on a specific object, i.e., face, we consider the facial gender recogni-tion as the next topic. First, we propose a multi-scale learned pattern (MSLP) basedface descriptor. The proposed feature learns multi-scale convolution templates via oneof algorithms like ICA, PCA or K-means, then encodes the images according to theresponse order of these templates, and finally forms a more compact and discrimi-native histogram feature for facial representation. Having witnessed the success oflinear representation based classifiers (LinearRCs) in face recognition and consideredthe different characteristics of data distribution between face recognition and genderrecognition, we systematically analyze the application of LinearRCs on facial gen-der recognition. Stemming from the idea of prototype generation, a LinearRC usingpartial least squares (i.e., LRC PLS) and its group based variant are proposed. The ex-tensive experiments on facial gender recognition show that the features using MSLPare better than previous hand-engineered features, that LRC PLS is more stable thanother LinearRCs while having relatively less prediction time, and that its group variantcan further boost the performance.
What’s more, in light of the success of prototype concept in gender recognition,the dissertation presents a LinearRC named multi-scale query-expanded collabora-tive representation based classifier with class-specific prototypes (QCRC CP) fromthe viewpoint of dictionary learning in order to deal with more complex object recog-nition (i.e., to deal with objects with multiple pose or view, more generic objects, etc.).We firstly expand a single query image into a query set by scaling, and then constructa query-dependent dictionary via canonical correlation analysis (CCA) between thequery set and the samples from a specific class. By integrating the prototype selectionbased methods and prototype generation based methods, the proposed method of dic-tionary learning generates query-related class-specific prototypes in a class-wise way, in which the implicit query-dependent data locality discards the outliers. In the end,multivariate collaborative representation based classification (CRC) using newly con-structed dictionary is employed to determine the identity according to the rule of mini-mum normalized residual (MNR). The experiments cover multi-pose face recognition,leaf species recognition, character recognition in natural scene and generic multi-viewobject recognition. The results demonstrate that QCRC CP can gain pleasing resultsand the proposed approach of dictionary learning is superior to previous methods viaprototype selection or generation. For instance, in contrast to other LinearRCs, QCR-C CP improves the recognition accuracy with an increase of over10%on characterrecognition.
Last but not least, considering a proper distance metric in scene recognition caneffectively reflect semantic distance between samples in high-dimensional space, wepropose a metric learning algorithm based on regularized linear discriminant anal-ysis (RLDA) to learn a full parameter matrix of Mahalanobis distance metric. It’sacknowledged that high-dimensional features with relatively limited samples mightintroduce huge computational cost and overfitting in conventional metric learning. Anovel framework is thus presented to resolve these problems. In our framework, theestimation of the parameter matrix is decomposed into that of a projection pool ma-trix and a non-negative diagonal selection matrix, which can significantly reduce thenumber of estimated parameters to a small scale. We first construct the projection vec-tors by tuning regularized LDA with different parameters. Based on side information,nonnegative l2-norm regularized least squares is then used to select and weigh the pro-jection vectors on the constructed training dataset. The training dataset is comprisedof the pairwise squared differences of the projected samples in the dissimilarity andsimilarity pair subsets. To maintain the balance of two subsets, a simple but effective s-trategy based on K nearest neighbors (KNN) is provided. It turns out that the proposedalgorithm can achieve higher recognition rates on two datasets for scene recognitionwhile retaining such high efficiency that obtains several-fold, and even dozens-foldacceleration compared with conventional metric learning methods.
引文
[1]章章毓晋.图像处理和分析技术[M].北京:清华大学出版社,2008.
[2]高隽,谢昭.图像理解理论与方法[M].北京:科学出版社,2009.
[3] FARRELL R, OZA O, ZHANG N, et al. Birdlets: Subordinate categorizationusing volumetric primitives and pose-normalized appearance[C]//IEEE Interna-tional Conference on Computer Vision. Barcelona: IEEE,2011:161–168.
[4] YANG S, BO L, WANG J, et al. Unsupervised template learning for fine-grained object recognition[C]//26th Annual Conference on Neural InformationProcessing Systems. Lake Tahoe: NIPS,2012:3131–3139.
[5] FARRELL R M. Learning visual patterns: Imposing order on objects, trajecto-ries and networks[D]. College Park: University of Maryland,2011.
[6] WAGNER A, WRIGHT J, GANESH A, et al. Toward a practical face recognitionsystem: Robust alignment and illumination by sparse representation[J]. IEEETransactions on Pattern Analysis and Machine Intelligence,2012,34(2):372–386.
[7] STARK M, KRAUSE J, PEPIK B, et al. Fine-grained categorization for3Dscene understanding[J]. International Journal of Robotics Research,2011,30(13):1543–1552.
[8] PRINCE S J D. Computer vision: Models, learning, and inference[M]. NewYork: Cambridge University Press,2012.
[9] OLIVA A, TORRALBA A. Modeling the shape of the scene: A holistic repre-sentation of the spatial envelope[J]. International Journal of Computer Vision,2001,42(3):145–175.
[10]陆福相.基于内容的场景和物体类别识别研究[D].上海:上海交通大学,2011.
[11] BARNARD K. Practical colour constancy[D]. Burnaby: Simon Fraser Univer-sity,1999.
[12] EBNER M. Color constancy[M]. England: Wiley,2007.
[13] CARDEI V C, FUNT B, BARNARD K. Estimating the scene illumination chro-maticity by using a neural network[J]. Journal of the Optical Society of AmericaA,2002,19(12):2374–2386.
[14] SWAIN M J, BALLARD D H. Color indexing[J]. International Journal of Com-puter Vision,1991,7(1):11–32.
[15] STRICKER M A, ORENGO M. Similarity of color images[C]//SPIE Conferenceon Storage and Retrieval for Image and Video Databases III. San Jose: SPIE,1995:381–392.
[16] SMITH J R, CHANG S F. Tools and techniques for color image re-trieval[C]//SPIE Conference on Storage and Retrieval for Still Image and VideoDatabases IV. San Jose: SPIE,1996:426–437.
[17] DALAL N, TRIGGS B. Histograms of oriented gradients for human de-tection[C]//IEEE International Conference on Computer Vision and PatternRecognition. San Diego: IEEE,2005:886–893.
[18] OJALA T, PIETIK A¨INEN M, HARWOOD D. A comparative study of texturemeasures with classification based on featured distributions[J]. Pattern Recog-nition,1996,29(1):51–59.
[19] AXELROD S, GOEL V, GOPINATH R, et al. Subspace constrained Gaussianmixture models for speech recognition[J]. IEEE Transactions on Speech andAudio Processing,2005,13(6):1144–1160.
[20] COATES A, CARPENTER B, CASE C, et al. Text detection and characterrecognition in scene images with unsupervised feature learning[C]//11th Inter-national Conference on Document Analysis and Recognition. Beijing: IEEE,2011:440–445.
[21] NETZER Y, WANG T, COATES A, et al. Reading digits in natural images withunsupervised feature learning[C]//25th Annual Conference on Neural Informa-tion Processing Systems Workshop. Granada: NIPS,2011.
[22] CORTES C, VAPNIK V. Support vector machine[J]. Machine Learning,1995,20(3):273–297.
[23] WRIGHT J, YANG A Y, GANESH A, et al. Robust face recognition via sparserepresentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelli-gence,2009,31(2):210–227.
[24] ZHANG L, YANG M, FENG X. Sparse representation or collaborative rep-resentation: Which helps face recognition?[C]//IEEE International Conferenceon Computer Vision. Barcelona: IEEE,2011:471–478.
[25]章鹏.多尺度特征检测:方法和应用研究[D].合肥:中国科学技术大学,2010.
[26]刘立.基于多尺度特征的图像匹配与目标定位研究[D].武汉:华中科技大学,2008.
[27] LINDEBERG T. Scale-space theory: A basic tool for analyzing structures atdifferent scales[J]. Journal of applied statistics,1994,21(1-2):225–270.
[28] BURT P, ADELSON E. The Laplacian pyramid as a compact image code[J].IEEE Transactions on Communications,1983,31(4):532–540.
[29] CHUI C K. Wavelets: A mathematical tool for signal analysis[M]. Philadelphia:SIAM,1997.
[30] LAZEBNIK S, SCHMID C, PONCE J. Beyond bags of features: Spatial pyra-mid matching for recognizing natural scene categories[C]//IEEE InternationalConference on Computer Vision and Pattern Recognition. New York: IEEE,2006:2169–2178.
[31] WU J, REHG J M. CENTRIST: A visual descriptor for scene categoriza-tion[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(8):1489–1501.
[32] SIAGIAN C, ITTI L. Gist: A mobile robotics application of context-basedvision in outdoor environment[C]//IEEE International Conference on ComputerVision and Pattern Recognition Workshops. San Diego: IEEE,2005:88.
[33] TIKHONOV A. Solution of incorrectly formulated problems and the regulariza-tion method[J]. Soviet Mathematics Doklady,1963,4:1035–1038.
[34] HOERL A E, KENNARD R W. Ridge regression: Biased estimation fornonorthogonal problems[J]. Technometrics,1970,12(1):55–67.
[35] TIBSHIRANI R. Regression shrinkage and selection via the lasso[J]. Journal ofthe Royal Statistical Society, Series B,1996,58(1):267–288.
[36] ZOU H, HASTIE T. Regularization and variable selection via the elastic net[J].Journal of the Royal Statistical Society: Series B,2005,67(2):301–320.
[37] YUAN M, LIN Y. Model selection and estimation in regression with groupedvariables[J]. Journal of the Royal Statistical Society: Series B,2006,68(1):49–67.
[38] MAJUMDAR A, WARD R. Robust classifiers for data reduced via random pro-jections[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cy-bernetics,2010,40(5):1359–1371.
[39] VAPNIK V N. The nature of statistical learning theory[M]. New York: Springer-Verlag,1999.
[40] HASTIE T J, TIBSHIRANI R J, FRIEDMAN J J H. The elements of statisticallearning[M]. New York: Springer-Verlag,2009.
[41] GIJSENIJ A, GEVERS T. Color constancy using natural image statistic-s[C]//IEEE International Conference on Computer Vision and Pattern Recog-nition. Minneapolis: IEEE,2007:1–8.
[42] HORDELY S D. Scene illuminant estimation: Past, present and future[J]. ColorResearch and Applicaiton,2006,31(4):303–314.
[43] SCHAEFER G, HORDLEY S, FINALAYSON G. A combined physical and s-tatistical approach to colour constancy[C]//IEEE International Conference onComputer Vision and Pattern Recognition. San Diego: IEEE,2005:148–153.
[44] LI B, XU D, LANG C Y. Colour constancy based on texture similarity fornatural images[J]. Coloration Technology,2009,125(6):328–333.
[45] GIJSENIJ A, GEVERS T. Color constancy using natural image statistics andscene semantics[J]. IEEE Transactions on Pattern Analysis and Machine Intel-ligence,2011,33(4):687–698.
[46] BIANCO S, GASPARINI F, SCHETTINI R. Consensus-based framework forilluminant chromaticity estimation[J]. Journal of Electronic Imaging,2008,17(2):023013–023013–9.
[47] PROVENZI E, GATTA C, FIERRO M, et al. A spatially variant white-patchand gray-world method for color image enhancement driven by local con-trast[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(10):1757–1770.
[48] CARDEI V C, FUNT B. Committee-based colour constancy[C]//7th ColorImaging Conference: Color Science, Systems and Applications. Scottsdale:IS&T,1999:311–313.
[49] BIANCO S, CIOCCA G, CUSANO C, et al. Improving color constancy usingindoor-outdoor image classification[J]. IEEE Transactions on Image Process-ing,2008,17(12):2381–2392.
[50] LAND E, MCCANN J. Lightness and retinex theory[J]. Journal of the OpticalSociety of America: A,1971,61(1):1–11.
[51] BUCHSBAUM G. A spatial processor model for object colour perception[J].Journal of The Franklin Institute,1980,310(1):1–26.
[52] FINLAYSON G D, TREZZI E. Shades of gray and colour constancy[C]//12thColor Imaging Conference: Color Science and Engineering Systems, Technolo-gies, and Applications. Scottsdale: IS&T,2004:37–41.
[53] VAN DE WEIJER J, GEVERS T. Color constancy based on the grey-edgehypothesis[C]//IEEE International Conference on Image Processing. Genova:IEEE,2005:722–725.
[54] VAN DE WEIJER J, SCHMID C, VERBEEK J. Using highlevel visual informa-tion for color constancy[C]//IEEE International Conference on Computer Vi-sion. Rio de Janeiro: IEEE,2007:1–8.
[55] VAN DE WEIJER J, GEVERS T, GIJSENIJ A. Edge-based color constancy[J].IEEE Transactions on Image Processing,2007,16(9):2207–2214.
[56] BIANCO S, CIOCCA G, CUSANO C, et al. Automatic color constancy algorith-m selection and combination[J]. Pattern Recognition,2010,43(3):695–705.
[57] CIUREA F, FUNT B V. A large image database for color constancy re-search[C]//11th Color Imaging Conference: Color Science and EngineeringSystems, Technology and Applications. Scottsdale: IS&T,2003:160–164.
[58] GEHLER P V, ROTHER C, BLAKE A, et al. Bayesian color constancy revisit-ed[C]//IEEE International Conference on Computer Vision and Pattern Recog-nition. Anchorage: IEEE,2008:1–8.
[59] GEUSEBROEK J M, SMEULDERS A W M. A six-stimulus theory for stochastictexture[J]. International Jornal of Computer Vision,2005,62(1–2):7–16.
[60] YANULEVSKAYA V, VAN GEMERT J C, K. ROTH A K H, et al. Emotion-al valence categorization using holistic image features[C]//IEEE InternationalConference on Image Processing. San Diego: IEEE,2008:101–104.
[61] GEUSEBROEK J M. Compact object descriptors from local colour invari-ant histograms[C]//British Machine Vision Conference. Edinburgh: BMVA,2006:1029–1038.
[62] VAN GEMERT J C, SNOEK C G M, VEENMAN C J, et al. Comparing compactcodebooks for visual categorization[J]. Computer Vision and Image Under-standing,2010,114(4):450–624.
[63] LU R, GIJSENIJ A, GEVERS T, et al. Color constancy using stage classifi-cation[C]//IEEE International Conference on Image Processing. Cairo: IEEE,2009:685–688.
[64] RUDERMAN D L, CRONIN T W, CHIAO C C. Statistics of cone responses tonatural images: Implications for visual coding[J]. Journal of the Optical Societyof America A,1998,15(8):2036–2045.
[65] HRUSTEMOVIC N, GUPTA M R. Multiresolutional regularization of locallinear regression over adaptive neighborhoods for color management[C]//IEEEInternational Conference on Image Processing. San Diego: IEEE,2008:497–500.
[66] PORIKLI F. Integral histogram: A fast way to extract histograms in carte-sian spaces[C]//IEEE International Conference on Computer Vision and PatternRecognition. San Diego: IEEE,2005:829–836.
[67] LOWE D G. Distinctive image features from scale-invariant keypoints[J]. Jour-nal of the Optical Society of America A,2004,60(2):91–110.
[68] BEKIOS-CALFA J, BUENAPOSADA J, BAUMELA L. Revisiting linear discrim-inant techniques in gender recognition[J]. IEEE Transactions on Pattern Anal-ysis and Machine Intelligence,2011,33(4):858–864.
[69] HUANG D, SHAN C, ARDABILIAN M, et al. Local binary patterns and its ap-plication to facial image analysis: A survey[J]. IEEE Transactions on Systems,Man, and Cybernetics, Part C: Applications and Reviews,2011,41(6):765–781.
[70] SHAN C. Learning local binary patterns for gender classification on real-worldface images[J]. Pattern Recognition Letters,2012,33(4):431–437.
[71] JABID T, KABIR M, CHAE O. Gender classification using local direction-al pattern (LDP)[C]//IEEE Ineternational Conference on Pattern Recognition.Istanbul: IEEE,2010:2162–2165.
[72] SCHIH H. Robust gender classification using a precise patch histogram[J]. Pat-tern Recognition,2013,46(2):519–528.
[73] KAWULOK M, WU J, HANCOCK E R. Supervised relevance maps for increas-ing the distinctiveness of facial images[J]. Pattern Recognition,2011.
[74] CHU W, HUANG C, CHEN C. Gender classification from unaligned facialimages using support subspaces[J]. Information Sciences,2013,221:98–109.
[75] VIOLA P, JONES M. Robust real-time face detection[J]. International Journalof Computer Vision,2004,57(2):137–154.
[76] MAKINEN E, RAISAMO R. An experimental comparison of gender classifica-tion methods[J]. Pattern Recognition Letters,2008,29(10):1544–1556.
[77] GUO G, DYER C, FU Y, et al. Is gender recognition affected by age?[C]//IEEEInternational Conference on Computer Vision Workshops. Kyoto: IEEE,2009:2032–2039.
[78] ALEXANDRE L. Gender recognition: A multiscale decision fusion approach[J].Pattern Recognition Letters,2010,31(11):1422–1427.
[79] HADID A, PIETIKAINEN M. Combining appearance and motion for face andgender recognition from videos[J]. Pattern Recognition,2009,42(11):2818–2827.
[80] FREUND Y, SCHAPIRE R. Experiments with a new boosting algorith-m[C]//Ineternational Conference on Machine Learning. Bari: IEEE,1996:148–156.
[81] GROSSO E, LAGORIO A, PULINA L, et al. Understanding critical factors inappearance-based gender categorization[C]//European Conference on Comput-er Vision. Firenze: Springer-Verlag,2012:280–289.
[82] LU H, HUANG Y, CHEN Y, et al. Automatic gender recognition based on pixel-pattern-based texture feature[J]. Journal of Real-Time Image Processing,2008,3(1–2):109–116.
[83] BALUJA S, ROWLEY H. Boosting sex identification performance[J]. Interna-tional Journal of Computer Vision,2007,71(1):111–119.
[84] WANG T, WU D, COATES A, et al. End-to-end text recognition with convolu-tional neural networks[C]//IEEE International Conference on Pattern Recogni-tion. Tsukuba: IEEE,2012:3304–3308.
[85] LE Q, KARPENKO A, NGIAM J, et al. ICA with reconstruction cost for ef-ficient overcomplete feature learning[C]//25th Annual Conference on NeuralInformation Processing Systems. Granada: NIPS,2011:1017–1025.
[86] XIAO Y, ZHU Z, WEI S, et al. Discriminative ICA model with reconstruc-tion constraint for image classification[C]//ACM International Conference onMultimedia. Nara: ACM,2012:929–932.
[87] OJALA T, PIETIKAINEN M, MAENPAA T. Multiresolution gray-scale and ro-tation invariant texture classification with local binary patterns[J]. Pattern Anal-ysis and Machine Intelligence, IEEE Transactions on,2002,24(7):971–987.
[88] LIAO S, ZHU X, LEI Z, et al. Learning multi-scale block local binary pat-terns for face recognition[C]//Advances in Biometrics. Seoul: Springer-Verlag,2007:828–837.
[89] COATES A, LEE H, NG A. An analysis of single-layer networks in unsu-pervised feature learning[C]//14th International Conference on Artificial Intel-ligence and Statistics. Fort Lauderdale: Microtome Publishing,2011:215–223.
[90] HYVARINEN A, OJA E. Independent component analysis: Algorithms andapplications[J]. Neural Networks,2000,13(4-5):411–430.
[91] WOLD S, MARTENS H, WOLD H. The multivariate calibration problemin chemistry solved by the PLS method[C]//International Conference on Ma-trix Pencils, Lecture Notes in Mathematics. Heidelberg: Springer-Verlag,1983:286–293.
[92] SCHWARTZ W, KEMBHAVI A, HARWOOD D, et al. Human detection usingpartial least squares analysis[C]//IEEE International Conference on ComputerVision. Kyoto: IEEE,2009:24–31.
[93] SCHWARTZ W R, GUO H, CHOI J, et al. Face identification using large featuresets[J]. IEEE Transactions on Image Processing,2012,21(4):2245–2255.
[94] ROSIPAL R, KRAMER N. Overview and recent advances in partial leastsquares[J]. Lecture Notes in Computer Science,2006:34–51.
[95] SAMSUDIN N, BRADLEY A. Nearest neighbour group-based classification[J].Pattern Recognition,2010,43(10):3458–3467.
[96] PHILLIPS P, MOON H, RIZVI S, et al. The FERET evaluation methodologyfor face recognition algorithms[J]. IEEE Transactions on Pattern Analysis andMachine Intelligence,2000,22(10):1090–1104.
[97] MINEAR M, PARK D. A lifespan database of adult facial stimuli[J]. BehaviorResearch Methods, Instruments,&Computers,2004,36(4):630–633.
[98] GALLAGHER A, CHEN T. Understanding images of groups of people[C]//IEEEComputer Society Conference on Computer Vision and Pattern RecognitionWorkshops. Miami: IEEE,2009:256–263.
[99] SHAN C. Learning local features for age estimation on real-Life faces[C]//ACMWorkshop on Multimodal Pervasive Video Analysis. Firenze: ACM,2010:23–28.
[100] DAGO-CASAS P, GONZALEZ-JIMENEZ D, LONG L, et al. Single-andcross-database benchmarks for gender classification under unconstrainedsettings[C]//IEEE International Conference on Computer Vision Workshops.Barcelona: IEEE,2011:2152–2159.
[101] MAKINEN E, RAISAMO R. Evaluation of gender classification methods withautomatically detected and aligned faces[J]. IEEE Transactions on PatternAnalysis and Machine Intelligence,2008,30(3):541–547.
[102] SHESHADRI K, DIVVALA S K. Exemplar driven character recognition in thewild[C]//British Machine Vision Conference. Surrey: BMVA,2012.
[103] WRIGHT J, MA Y, MAIRAL J, et al. Sparse representation for computer visionand pattern recognition[J]. Proceedings of IEEE,2010:1031–1044.
[104] LU C, MIN H, GUI J, et al. Face recognition via weighted sparse repre-sentation[J]. Journal of Visual Communicaiton Image Representation,2013,24(2):111–116.
[105] YANG J, ZHANG L, XU Y, et al. Beyond sparsity: The role of L1-optimizer inpattern classification[J]. Pattern Recognition,2012,45(3):1104–1118.
[106] HE R, ZHENG W, HU B. Maximum correntropy criterion for robust face recog-nition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(8):1561–1576.
[107] ELHAMIFAR E, VIDAL R. Robust classification using structured sparse repre-sentation[C]//IEEE Conference on Computer Vision and Pattern Recognition.Colorado Springs: IEEE,2011:1873–1879.
[108] YANG M, ZHANG L, SIMON S C K, et al. Gabor feature based robust represen-tation and classification for face recognition with Gabor occlusion dictionary[J].Pattern Recognition,2013,46(7):1865–1878.
[109] HUANG J, YANG M. Fast sparse representation with prototypes[C]//IEEEInternational Conference on Computer Vision and Pattern Recognition. SanFrancisco: IEEE,2010:3618–3625.
[110] YANG M, ZHANG L, YANG J, et al. Metaface learning for sparse representationbased face recognition[C]//. Hong Kong: IEEE,2010:1601–1604.
[111] WEI C, CHAO Y, YEH Y, et al. Locality-sensitive dictionary learning for sparserepresentation based classification[J]. Pattern Recognition,2013,46(5):1277–1287.
[112] XU J, YANG J. Mean representation based classifier with its applications[J].Electronics Letters,2011,47(18):1024–1026.
[113] XU Y, ZHU Q. A simple and fast representation-based face recognitionmethod[J]. Neural Computing and Applications,2012.
[114] HE R, ZHENG W, HU B, et al. Two-stage nonnegative sparse representationfor large-scale face recognition[J]. IEEE Transactions on Neural Networks andLearning System,2013,24(1):35–46.
[115] XU Y, ZHANG D, YANG J, et al. A two-phase test sample sparse representationmethod for use with face recognition[J]. IEEE Transactions on Circuits andSystems for Video Technology,2011,21(9):1255–1262.
[116] XU Y, ZHU X, LI Z, et al. Using the original and’symmetrical face’ trainingsamples to perform representation based two-step face recognition[J]. PatternRecognition,2013,46(4):1151–1158.
[117] WU X, WU N. Distance approximation for two-phase test sample representa-tion in face recognition[J]. Neural Computing and Applications,2013.
[118] CHEN Y, PATEL V, SHEKHAR S, et al. Video-based face recognition via join-t sparse representation[C]//IEEE International Conference on Automatic Faceand Gesture Recognition. Shanghai: IEEE,2013.
[119] ZHANG H, ZHANG Y, HUANG T S. Pose-robust face recognition via sparserepresentation[J]. Pattern Recognition,2013,46(5):1511–1521.
[120] YUAN X T, YAN S. Visual classification with multi-task joint sparse represen-tation[C]//IEEE Conference on Computer Vision and Pattern Recognition. SanFrancisco: IEEE,2010:3493–3500.
[121] YANG M, ZHANG L, ZHANG D, et al. Relaxed collaborative representationfor pattern classification[C]//IEEE Conference on Computer Vision and PatternRecognition. Providence: IEEE,2012:2224–2231.
[122] HU Y, MIAN A S, OWENS R. Face recognition using sparse approximatednearest points between image sets[J]. IEEE Transactions on Pattern Analysisand Machine Intelligence,2012,34(10):1992–2004.
[123] KIM T, KITTLER J, CIPOLLA R. Discriminative learning and recognition ofimage set classes using canonical correlations[J]. IEEE Transactions on PatternAnalysis and Machine Intelligence,2007,29(6):1005–1018.
[124] TRIGUERO I, DERRAC J, GARCIA S, et al. A taxonomy and experimental studyon prototype generation for nearest neighbor classification[J]. IEEE Transac-tions on Systems, Man, and Cybernetics, Part C: Applications and Reviews,2012,42(1):86–100.
[125] AHARON M, ELAD M, BRUCKSTEIN A. K-SVD: An algorithm for designingof overcomplete dictionaries for sparse representation[J]. IEEE Transactions onSignal Processing,2006,54(11):4311–4322.
[126] ZHOU L, ZHOU Z, HU D. Scene classification using a multi-resolution bag-of-features model[J]. Pattern Recognition,2013,46(1):424–433.
[127] HASSNER T, MAYZELS V, ZELNIK-MANOR L. On SIFTs and their s-cales[C]//IEEE Conference on Computer Vision and Pattern Recognition. Prov-idence: IEEE,2012:1522–1528.
[128] LI J, LU B L. An adaptive image Euclidean distance[J]. Pattern Recognition,2009,42(3):349–357.
[129] WANG L, ZHANG Y, FENG J. On the Euclidean distance of images[J]. IEEETransactions on Pattern Analysis and Machine Intelligence,2005,27(8).
[130] YU K, ZHANG T, GONG Y. Nonlinear learning using local coordinate cod-ing[C]//23rd Annual Conference on Neural Information Processing Systems.Vancouver: NIPS,2009.
[131] SU Y, FU Y, GAO X B, et al. Discriminant learning through multiple principalangles for visual recognition[J]. IEEE Transactions on Image Processing,2012,21(3):1381–1390.
[132] BOUTELL M, BROWN C, LUO J. Review of the state of the art in semanticscene classificaiton[R]. Rochester: University of Rochester,2002.
[133] MENG X, WANG Z, WU L. Building global image features for scene recogni-tion[J]. Pattern Recognition,2012,45(1):373–380.
[134] VAILAYA A, FIGUEIREDO M, JAIN A, et al. Content-based hierachical classi-fication of vacation images[C]//IEEE International Conference on MultimediaComputing and Systems. Florence: IEEE,1999:518–523.
[135] GOKALP D, AKSOY S. Using bag-of-regions representations[C]//IEEE Inter-national Conference on Computer Vision and Pattern Recognition. Minneapo-lis: IEEE,2007:1–8.
[136] RAMANAN D, BAKER S. Local distance functions: A taxonomy, new algo-rithms and an evaluation[J]. IEEE Transactions on Pattern Analysis and Ma-chine Intelligence,2011,33(4):794–806.
[137] XING E, NG A, JORDAN M, et al. Distance metric learning, with applica-tion to clustering with side-information[C]//17th Annual Conference on NeuralInformation Processing Systems. Whistler: NIPS,2003:505–512.
[138] XIANG S, NIE F, ZHANG C. Learning a Mahalanobis distance metric for dataclustering and classification[J]. Pattern Recognition,2008,41(12):3600–3612.
[139] WEINBERGER K, SAUL L. Distance metric learning for large margin near-est neighbor classification[J]. Journal of Machine Learning Research,2009,10:207–244.
[140] DAVIS J, KULIS B, JAIN P, et al. Information-theoretic metric learn-ing[C]//International Conference on Machine Learning. Corvallis: ACM,2007:209–216.
[141] GOLDBERGER J, ROWEIS S, HINTON G, et al. Neighbourhood componentsanalysis[C]//18th Annual Conference on Neural Information Processing Sys-tems. Vancouver: NIPS,2004:513–520.
[142] YANG L, JIN R, MUMMERT L, et al. A boosting framework for visuality-preserving distance metric learning and its application to medical image re-trieval[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2010,32(1):30–44.
[143] SHEN C, KIM J, WANG L. Scalable large-margin Mahalanobis distance metriclearning[J]. IEEE Transactions on Neural Networks,2010,21(9):1524–1530.
[144] WOZNICA A, KALOUSIS A. A new framework for dissimilarity and simi-larity learning[C]//Pacific-Asia Conference on Knowledge Discovery and DataMining. Hyderabad: Springer-Verlag,2010:386–397.
[145] YU J, AMORES J, SEBE N, et al. Distance learning for similarity estima-tion[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(3):451–462.
[146] FUKUNAGA K. Introduction to Statistical Pattern Recognition[M]. New York:Academic Press,1999.
[147] GOLUB G H, LOAN C V. Matrix Computation[M]. Baltimore: The JohnsHopking University Press,1996.
[148] LI L, LI F. What, where, and who? Classifying events by scene and objectrecognition[C]//IEEE Internatioanl Conference on Computer Vision. Rio deJaneiro: IEEE,2007:1–8.
[149] ZHU J, HOI S, LYU M, et al. Near-duplicate keyframe retrieval by nonrigid im-age matching[C]//ACM International Conference on Multimedia. Vancouver:ACM,2008:41–50.
[2]高隽,谢昭.图像理解理论与方法[M].北京:科学出版社,2009.
[3] FARRELL R, OZA O, ZHANG N, et al. Birdlets: Subordinate categorizationusing volumetric primitives and pose-normalized appearance[C]//IEEE Interna-tional Conference on Computer Vision. Barcelona: IEEE,2011:161–168.
[4] YANG S, BO L, WANG J, et al. Unsupervised template learning for fine-grained object recognition[C]//26th Annual Conference on Neural InformationProcessing Systems. Lake Tahoe: NIPS,2012:3131–3139.
[5] FARRELL R M. Learning visual patterns: Imposing order on objects, trajecto-ries and networks[D]. College Park: University of Maryland,2011.
[6] WAGNER A, WRIGHT J, GANESH A, et al. Toward a practical face recognitionsystem: Robust alignment and illumination by sparse representation[J]. IEEETransactions on Pattern Analysis and Machine Intelligence,2012,34(2):372–386.
[7] STARK M, KRAUSE J, PEPIK B, et al. Fine-grained categorization for3Dscene understanding[J]. International Journal of Robotics Research,2011,30(13):1543–1552.
[8] PRINCE S J D. Computer vision: Models, learning, and inference[M]. NewYork: Cambridge University Press,2012.
[9] OLIVA A, TORRALBA A. Modeling the shape of the scene: A holistic repre-sentation of the spatial envelope[J]. International Journal of Computer Vision,2001,42(3):145–175.
[10]陆福相.基于内容的场景和物体类别识别研究[D].上海:上海交通大学,2011.
[11] BARNARD K. Practical colour constancy[D]. Burnaby: Simon Fraser Univer-sity,1999.
[12] EBNER M. Color constancy[M]. England: Wiley,2007.
[13] CARDEI V C, FUNT B, BARNARD K. Estimating the scene illumination chro-maticity by using a neural network[J]. Journal of the Optical Society of AmericaA,2002,19(12):2374–2386.
[14] SWAIN M J, BALLARD D H. Color indexing[J]. International Journal of Com-puter Vision,1991,7(1):11–32.
[15] STRICKER M A, ORENGO M. Similarity of color images[C]//SPIE Conferenceon Storage and Retrieval for Image and Video Databases III. San Jose: SPIE,1995:381–392.
[16] SMITH J R, CHANG S F. Tools and techniques for color image re-trieval[C]//SPIE Conference on Storage and Retrieval for Still Image and VideoDatabases IV. San Jose: SPIE,1996:426–437.
[17] DALAL N, TRIGGS B. Histograms of oriented gradients for human de-tection[C]//IEEE International Conference on Computer Vision and PatternRecognition. San Diego: IEEE,2005:886–893.
[18] OJALA T, PIETIK A¨INEN M, HARWOOD D. A comparative study of texturemeasures with classification based on featured distributions[J]. Pattern Recog-nition,1996,29(1):51–59.
[19] AXELROD S, GOEL V, GOPINATH R, et al. Subspace constrained Gaussianmixture models for speech recognition[J]. IEEE Transactions on Speech andAudio Processing,2005,13(6):1144–1160.
[20] COATES A, CARPENTER B, CASE C, et al. Text detection and characterrecognition in scene images with unsupervised feature learning[C]//11th Inter-national Conference on Document Analysis and Recognition. Beijing: IEEE,2011:440–445.
[21] NETZER Y, WANG T, COATES A, et al. Reading digits in natural images withunsupervised feature learning[C]//25th Annual Conference on Neural Informa-tion Processing Systems Workshop. Granada: NIPS,2011.
[22] CORTES C, VAPNIK V. Support vector machine[J]. Machine Learning,1995,20(3):273–297.
[23] WRIGHT J, YANG A Y, GANESH A, et al. Robust face recognition via sparserepresentation[J]. IEEE Transactions on Pattern Analysis and Machine Intelli-gence,2009,31(2):210–227.
[24] ZHANG L, YANG M, FENG X. Sparse representation or collaborative rep-resentation: Which helps face recognition?[C]//IEEE International Conferenceon Computer Vision. Barcelona: IEEE,2011:471–478.
[25]章鹏.多尺度特征检测:方法和应用研究[D].合肥:中国科学技术大学,2010.
[26]刘立.基于多尺度特征的图像匹配与目标定位研究[D].武汉:华中科技大学,2008.
[27] LINDEBERG T. Scale-space theory: A basic tool for analyzing structures atdifferent scales[J]. Journal of applied statistics,1994,21(1-2):225–270.
[28] BURT P, ADELSON E. The Laplacian pyramid as a compact image code[J].IEEE Transactions on Communications,1983,31(4):532–540.
[29] CHUI C K. Wavelets: A mathematical tool for signal analysis[M]. Philadelphia:SIAM,1997.
[30] LAZEBNIK S, SCHMID C, PONCE J. Beyond bags of features: Spatial pyra-mid matching for recognizing natural scene categories[C]//IEEE InternationalConference on Computer Vision and Pattern Recognition. New York: IEEE,2006:2169–2178.
[31] WU J, REHG J M. CENTRIST: A visual descriptor for scene categoriza-tion[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(8):1489–1501.
[32] SIAGIAN C, ITTI L. Gist: A mobile robotics application of context-basedvision in outdoor environment[C]//IEEE International Conference on ComputerVision and Pattern Recognition Workshops. San Diego: IEEE,2005:88.
[33] TIKHONOV A. Solution of incorrectly formulated problems and the regulariza-tion method[J]. Soviet Mathematics Doklady,1963,4:1035–1038.
[34] HOERL A E, KENNARD R W. Ridge regression: Biased estimation fornonorthogonal problems[J]. Technometrics,1970,12(1):55–67.
[35] TIBSHIRANI R. Regression shrinkage and selection via the lasso[J]. Journal ofthe Royal Statistical Society, Series B,1996,58(1):267–288.
[36] ZOU H, HASTIE T. Regularization and variable selection via the elastic net[J].Journal of the Royal Statistical Society: Series B,2005,67(2):301–320.
[37] YUAN M, LIN Y. Model selection and estimation in regression with groupedvariables[J]. Journal of the Royal Statistical Society: Series B,2006,68(1):49–67.
[38] MAJUMDAR A, WARD R. Robust classifiers for data reduced via random pro-jections[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cy-bernetics,2010,40(5):1359–1371.
[39] VAPNIK V N. The nature of statistical learning theory[M]. New York: Springer-Verlag,1999.
[40] HASTIE T J, TIBSHIRANI R J, FRIEDMAN J J H. The elements of statisticallearning[M]. New York: Springer-Verlag,2009.
[41] GIJSENIJ A, GEVERS T. Color constancy using natural image statistic-s[C]//IEEE International Conference on Computer Vision and Pattern Recog-nition. Minneapolis: IEEE,2007:1–8.
[42] HORDELY S D. Scene illuminant estimation: Past, present and future[J]. ColorResearch and Applicaiton,2006,31(4):303–314.
[43] SCHAEFER G, HORDLEY S, FINALAYSON G. A combined physical and s-tatistical approach to colour constancy[C]//IEEE International Conference onComputer Vision and Pattern Recognition. San Diego: IEEE,2005:148–153.
[44] LI B, XU D, LANG C Y. Colour constancy based on texture similarity fornatural images[J]. Coloration Technology,2009,125(6):328–333.
[45] GIJSENIJ A, GEVERS T. Color constancy using natural image statistics andscene semantics[J]. IEEE Transactions on Pattern Analysis and Machine Intel-ligence,2011,33(4):687–698.
[46] BIANCO S, GASPARINI F, SCHETTINI R. Consensus-based framework forilluminant chromaticity estimation[J]. Journal of Electronic Imaging,2008,17(2):023013–023013–9.
[47] PROVENZI E, GATTA C, FIERRO M, et al. A spatially variant white-patchand gray-world method for color image enhancement driven by local con-trast[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(10):1757–1770.
[48] CARDEI V C, FUNT B. Committee-based colour constancy[C]//7th ColorImaging Conference: Color Science, Systems and Applications. Scottsdale:IS&T,1999:311–313.
[49] BIANCO S, CIOCCA G, CUSANO C, et al. Improving color constancy usingindoor-outdoor image classification[J]. IEEE Transactions on Image Process-ing,2008,17(12):2381–2392.
[50] LAND E, MCCANN J. Lightness and retinex theory[J]. Journal of the OpticalSociety of America: A,1971,61(1):1–11.
[51] BUCHSBAUM G. A spatial processor model for object colour perception[J].Journal of The Franklin Institute,1980,310(1):1–26.
[52] FINLAYSON G D, TREZZI E. Shades of gray and colour constancy[C]//12thColor Imaging Conference: Color Science and Engineering Systems, Technolo-gies, and Applications. Scottsdale: IS&T,2004:37–41.
[53] VAN DE WEIJER J, GEVERS T. Color constancy based on the grey-edgehypothesis[C]//IEEE International Conference on Image Processing. Genova:IEEE,2005:722–725.
[54] VAN DE WEIJER J, SCHMID C, VERBEEK J. Using highlevel visual informa-tion for color constancy[C]//IEEE International Conference on Computer Vi-sion. Rio de Janeiro: IEEE,2007:1–8.
[55] VAN DE WEIJER J, GEVERS T, GIJSENIJ A. Edge-based color constancy[J].IEEE Transactions on Image Processing,2007,16(9):2207–2214.
[56] BIANCO S, CIOCCA G, CUSANO C, et al. Automatic color constancy algorith-m selection and combination[J]. Pattern Recognition,2010,43(3):695–705.
[57] CIUREA F, FUNT B V. A large image database for color constancy re-search[C]//11th Color Imaging Conference: Color Science and EngineeringSystems, Technology and Applications. Scottsdale: IS&T,2003:160–164.
[58] GEHLER P V, ROTHER C, BLAKE A, et al. Bayesian color constancy revisit-ed[C]//IEEE International Conference on Computer Vision and Pattern Recog-nition. Anchorage: IEEE,2008:1–8.
[59] GEUSEBROEK J M, SMEULDERS A W M. A six-stimulus theory for stochastictexture[J]. International Jornal of Computer Vision,2005,62(1–2):7–16.
[60] YANULEVSKAYA V, VAN GEMERT J C, K. ROTH A K H, et al. Emotion-al valence categorization using holistic image features[C]//IEEE InternationalConference on Image Processing. San Diego: IEEE,2008:101–104.
[61] GEUSEBROEK J M. Compact object descriptors from local colour invari-ant histograms[C]//British Machine Vision Conference. Edinburgh: BMVA,2006:1029–1038.
[62] VAN GEMERT J C, SNOEK C G M, VEENMAN C J, et al. Comparing compactcodebooks for visual categorization[J]. Computer Vision and Image Under-standing,2010,114(4):450–624.
[63] LU R, GIJSENIJ A, GEVERS T, et al. Color constancy using stage classifi-cation[C]//IEEE International Conference on Image Processing. Cairo: IEEE,2009:685–688.
[64] RUDERMAN D L, CRONIN T W, CHIAO C C. Statistics of cone responses tonatural images: Implications for visual coding[J]. Journal of the Optical Societyof America A,1998,15(8):2036–2045.
[65] HRUSTEMOVIC N, GUPTA M R. Multiresolutional regularization of locallinear regression over adaptive neighborhoods for color management[C]//IEEEInternational Conference on Image Processing. San Diego: IEEE,2008:497–500.
[66] PORIKLI F. Integral histogram: A fast way to extract histograms in carte-sian spaces[C]//IEEE International Conference on Computer Vision and PatternRecognition. San Diego: IEEE,2005:829–836.
[67] LOWE D G. Distinctive image features from scale-invariant keypoints[J]. Jour-nal of the Optical Society of America A,2004,60(2):91–110.
[68] BEKIOS-CALFA J, BUENAPOSADA J, BAUMELA L. Revisiting linear discrim-inant techniques in gender recognition[J]. IEEE Transactions on Pattern Anal-ysis and Machine Intelligence,2011,33(4):858–864.
[69] HUANG D, SHAN C, ARDABILIAN M, et al. Local binary patterns and its ap-plication to facial image analysis: A survey[J]. IEEE Transactions on Systems,Man, and Cybernetics, Part C: Applications and Reviews,2011,41(6):765–781.
[70] SHAN C. Learning local binary patterns for gender classification on real-worldface images[J]. Pattern Recognition Letters,2012,33(4):431–437.
[71] JABID T, KABIR M, CHAE O. Gender classification using local direction-al pattern (LDP)[C]//IEEE Ineternational Conference on Pattern Recognition.Istanbul: IEEE,2010:2162–2165.
[72] SCHIH H. Robust gender classification using a precise patch histogram[J]. Pat-tern Recognition,2013,46(2):519–528.
[73] KAWULOK M, WU J, HANCOCK E R. Supervised relevance maps for increas-ing the distinctiveness of facial images[J]. Pattern Recognition,2011.
[74] CHU W, HUANG C, CHEN C. Gender classification from unaligned facialimages using support subspaces[J]. Information Sciences,2013,221:98–109.
[75] VIOLA P, JONES M. Robust real-time face detection[J]. International Journalof Computer Vision,2004,57(2):137–154.
[76] MAKINEN E, RAISAMO R. An experimental comparison of gender classifica-tion methods[J]. Pattern Recognition Letters,2008,29(10):1544–1556.
[77] GUO G, DYER C, FU Y, et al. Is gender recognition affected by age?[C]//IEEEInternational Conference on Computer Vision Workshops. Kyoto: IEEE,2009:2032–2039.
[78] ALEXANDRE L. Gender recognition: A multiscale decision fusion approach[J].Pattern Recognition Letters,2010,31(11):1422–1427.
[79] HADID A, PIETIKAINEN M. Combining appearance and motion for face andgender recognition from videos[J]. Pattern Recognition,2009,42(11):2818–2827.
[80] FREUND Y, SCHAPIRE R. Experiments with a new boosting algorith-m[C]//Ineternational Conference on Machine Learning. Bari: IEEE,1996:148–156.
[81] GROSSO E, LAGORIO A, PULINA L, et al. Understanding critical factors inappearance-based gender categorization[C]//European Conference on Comput-er Vision. Firenze: Springer-Verlag,2012:280–289.
[82] LU H, HUANG Y, CHEN Y, et al. Automatic gender recognition based on pixel-pattern-based texture feature[J]. Journal of Real-Time Image Processing,2008,3(1–2):109–116.
[83] BALUJA S, ROWLEY H. Boosting sex identification performance[J]. Interna-tional Journal of Computer Vision,2007,71(1):111–119.
[84] WANG T, WU D, COATES A, et al. End-to-end text recognition with convolu-tional neural networks[C]//IEEE International Conference on Pattern Recogni-tion. Tsukuba: IEEE,2012:3304–3308.
[85] LE Q, KARPENKO A, NGIAM J, et al. ICA with reconstruction cost for ef-ficient overcomplete feature learning[C]//25th Annual Conference on NeuralInformation Processing Systems. Granada: NIPS,2011:1017–1025.
[86] XIAO Y, ZHU Z, WEI S, et al. Discriminative ICA model with reconstruc-tion constraint for image classification[C]//ACM International Conference onMultimedia. Nara: ACM,2012:929–932.
[87] OJALA T, PIETIKAINEN M, MAENPAA T. Multiresolution gray-scale and ro-tation invariant texture classification with local binary patterns[J]. Pattern Anal-ysis and Machine Intelligence, IEEE Transactions on,2002,24(7):971–987.
[88] LIAO S, ZHU X, LEI Z, et al. Learning multi-scale block local binary pat-terns for face recognition[C]//Advances in Biometrics. Seoul: Springer-Verlag,2007:828–837.
[89] COATES A, LEE H, NG A. An analysis of single-layer networks in unsu-pervised feature learning[C]//14th International Conference on Artificial Intel-ligence and Statistics. Fort Lauderdale: Microtome Publishing,2011:215–223.
[90] HYVARINEN A, OJA E. Independent component analysis: Algorithms andapplications[J]. Neural Networks,2000,13(4-5):411–430.
[91] WOLD S, MARTENS H, WOLD H. The multivariate calibration problemin chemistry solved by the PLS method[C]//International Conference on Ma-trix Pencils, Lecture Notes in Mathematics. Heidelberg: Springer-Verlag,1983:286–293.
[92] SCHWARTZ W, KEMBHAVI A, HARWOOD D, et al. Human detection usingpartial least squares analysis[C]//IEEE International Conference on ComputerVision. Kyoto: IEEE,2009:24–31.
[93] SCHWARTZ W R, GUO H, CHOI J, et al. Face identification using large featuresets[J]. IEEE Transactions on Image Processing,2012,21(4):2245–2255.
[94] ROSIPAL R, KRAMER N. Overview and recent advances in partial leastsquares[J]. Lecture Notes in Computer Science,2006:34–51.
[95] SAMSUDIN N, BRADLEY A. Nearest neighbour group-based classification[J].Pattern Recognition,2010,43(10):3458–3467.
[96] PHILLIPS P, MOON H, RIZVI S, et al. The FERET evaluation methodologyfor face recognition algorithms[J]. IEEE Transactions on Pattern Analysis andMachine Intelligence,2000,22(10):1090–1104.
[97] MINEAR M, PARK D. A lifespan database of adult facial stimuli[J]. BehaviorResearch Methods, Instruments,&Computers,2004,36(4):630–633.
[98] GALLAGHER A, CHEN T. Understanding images of groups of people[C]//IEEEComputer Society Conference on Computer Vision and Pattern RecognitionWorkshops. Miami: IEEE,2009:256–263.
[99] SHAN C. Learning local features for age estimation on real-Life faces[C]//ACMWorkshop on Multimodal Pervasive Video Analysis. Firenze: ACM,2010:23–28.
[100] DAGO-CASAS P, GONZALEZ-JIMENEZ D, LONG L, et al. Single-andcross-database benchmarks for gender classification under unconstrainedsettings[C]//IEEE International Conference on Computer Vision Workshops.Barcelona: IEEE,2011:2152–2159.
[101] MAKINEN E, RAISAMO R. Evaluation of gender classification methods withautomatically detected and aligned faces[J]. IEEE Transactions on PatternAnalysis and Machine Intelligence,2008,30(3):541–547.
[102] SHESHADRI K, DIVVALA S K. Exemplar driven character recognition in thewild[C]//British Machine Vision Conference. Surrey: BMVA,2012.
[103] WRIGHT J, MA Y, MAIRAL J, et al. Sparse representation for computer visionand pattern recognition[J]. Proceedings of IEEE,2010:1031–1044.
[104] LU C, MIN H, GUI J, et al. Face recognition via weighted sparse repre-sentation[J]. Journal of Visual Communicaiton Image Representation,2013,24(2):111–116.
[105] YANG J, ZHANG L, XU Y, et al. Beyond sparsity: The role of L1-optimizer inpattern classification[J]. Pattern Recognition,2012,45(3):1104–1118.
[106] HE R, ZHENG W, HU B. Maximum correntropy criterion for robust face recog-nition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(8):1561–1576.
[107] ELHAMIFAR E, VIDAL R. Robust classification using structured sparse repre-sentation[C]//IEEE Conference on Computer Vision and Pattern Recognition.Colorado Springs: IEEE,2011:1873–1879.
[108] YANG M, ZHANG L, SIMON S C K, et al. Gabor feature based robust represen-tation and classification for face recognition with Gabor occlusion dictionary[J].Pattern Recognition,2013,46(7):1865–1878.
[109] HUANG J, YANG M. Fast sparse representation with prototypes[C]//IEEEInternational Conference on Computer Vision and Pattern Recognition. SanFrancisco: IEEE,2010:3618–3625.
[110] YANG M, ZHANG L, YANG J, et al. Metaface learning for sparse representationbased face recognition[C]//. Hong Kong: IEEE,2010:1601–1604.
[111] WEI C, CHAO Y, YEH Y, et al. Locality-sensitive dictionary learning for sparserepresentation based classification[J]. Pattern Recognition,2013,46(5):1277–1287.
[112] XU J, YANG J. Mean representation based classifier with its applications[J].Electronics Letters,2011,47(18):1024–1026.
[113] XU Y, ZHU Q. A simple and fast representation-based face recognitionmethod[J]. Neural Computing and Applications,2012.
[114] HE R, ZHENG W, HU B, et al. Two-stage nonnegative sparse representationfor large-scale face recognition[J]. IEEE Transactions on Neural Networks andLearning System,2013,24(1):35–46.
[115] XU Y, ZHANG D, YANG J, et al. A two-phase test sample sparse representationmethod for use with face recognition[J]. IEEE Transactions on Circuits andSystems for Video Technology,2011,21(9):1255–1262.
[116] XU Y, ZHU X, LI Z, et al. Using the original and’symmetrical face’ trainingsamples to perform representation based two-step face recognition[J]. PatternRecognition,2013,46(4):1151–1158.
[117] WU X, WU N. Distance approximation for two-phase test sample representa-tion in face recognition[J]. Neural Computing and Applications,2013.
[118] CHEN Y, PATEL V, SHEKHAR S, et al. Video-based face recognition via join-t sparse representation[C]//IEEE International Conference on Automatic Faceand Gesture Recognition. Shanghai: IEEE,2013.
[119] ZHANG H, ZHANG Y, HUANG T S. Pose-robust face recognition via sparserepresentation[J]. Pattern Recognition,2013,46(5):1511–1521.
[120] YUAN X T, YAN S. Visual classification with multi-task joint sparse represen-tation[C]//IEEE Conference on Computer Vision and Pattern Recognition. SanFrancisco: IEEE,2010:3493–3500.
[121] YANG M, ZHANG L, ZHANG D, et al. Relaxed collaborative representationfor pattern classification[C]//IEEE Conference on Computer Vision and PatternRecognition. Providence: IEEE,2012:2224–2231.
[122] HU Y, MIAN A S, OWENS R. Face recognition using sparse approximatednearest points between image sets[J]. IEEE Transactions on Pattern Analysisand Machine Intelligence,2012,34(10):1992–2004.
[123] KIM T, KITTLER J, CIPOLLA R. Discriminative learning and recognition ofimage set classes using canonical correlations[J]. IEEE Transactions on PatternAnalysis and Machine Intelligence,2007,29(6):1005–1018.
[124] TRIGUERO I, DERRAC J, GARCIA S, et al. A taxonomy and experimental studyon prototype generation for nearest neighbor classification[J]. IEEE Transac-tions on Systems, Man, and Cybernetics, Part C: Applications and Reviews,2012,42(1):86–100.
[125] AHARON M, ELAD M, BRUCKSTEIN A. K-SVD: An algorithm for designingof overcomplete dictionaries for sparse representation[J]. IEEE Transactions onSignal Processing,2006,54(11):4311–4322.
[126] ZHOU L, ZHOU Z, HU D. Scene classification using a multi-resolution bag-of-features model[J]. Pattern Recognition,2013,46(1):424–433.
[127] HASSNER T, MAYZELS V, ZELNIK-MANOR L. On SIFTs and their s-cales[C]//IEEE Conference on Computer Vision and Pattern Recognition. Prov-idence: IEEE,2012:1522–1528.
[128] LI J, LU B L. An adaptive image Euclidean distance[J]. Pattern Recognition,2009,42(3):349–357.
[129] WANG L, ZHANG Y, FENG J. On the Euclidean distance of images[J]. IEEETransactions on Pattern Analysis and Machine Intelligence,2005,27(8).
[130] YU K, ZHANG T, GONG Y. Nonlinear learning using local coordinate cod-ing[C]//23rd Annual Conference on Neural Information Processing Systems.Vancouver: NIPS,2009.
[131] SU Y, FU Y, GAO X B, et al. Discriminant learning through multiple principalangles for visual recognition[J]. IEEE Transactions on Image Processing,2012,21(3):1381–1390.
[132] BOUTELL M, BROWN C, LUO J. Review of the state of the art in semanticscene classificaiton[R]. Rochester: University of Rochester,2002.
[133] MENG X, WANG Z, WU L. Building global image features for scene recogni-tion[J]. Pattern Recognition,2012,45(1):373–380.
[134] VAILAYA A, FIGUEIREDO M, JAIN A, et al. Content-based hierachical classi-fication of vacation images[C]//IEEE International Conference on MultimediaComputing and Systems. Florence: IEEE,1999:518–523.
[135] GOKALP D, AKSOY S. Using bag-of-regions representations[C]//IEEE Inter-national Conference on Computer Vision and Pattern Recognition. Minneapo-lis: IEEE,2007:1–8.
[136] RAMANAN D, BAKER S. Local distance functions: A taxonomy, new algo-rithms and an evaluation[J]. IEEE Transactions on Pattern Analysis and Ma-chine Intelligence,2011,33(4):794–806.
[137] XING E, NG A, JORDAN M, et al. Distance metric learning, with applica-tion to clustering with side-information[C]//17th Annual Conference on NeuralInformation Processing Systems. Whistler: NIPS,2003:505–512.
[138] XIANG S, NIE F, ZHANG C. Learning a Mahalanobis distance metric for dataclustering and classification[J]. Pattern Recognition,2008,41(12):3600–3612.
[139] WEINBERGER K, SAUL L. Distance metric learning for large margin near-est neighbor classification[J]. Journal of Machine Learning Research,2009,10:207–244.
[140] DAVIS J, KULIS B, JAIN P, et al. Information-theoretic metric learn-ing[C]//International Conference on Machine Learning. Corvallis: ACM,2007:209–216.
[141] GOLDBERGER J, ROWEIS S, HINTON G, et al. Neighbourhood componentsanalysis[C]//18th Annual Conference on Neural Information Processing Sys-tems. Vancouver: NIPS,2004:513–520.
[142] YANG L, JIN R, MUMMERT L, et al. A boosting framework for visuality-preserving distance metric learning and its application to medical image re-trieval[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2010,32(1):30–44.
[143] SHEN C, KIM J, WANG L. Scalable large-margin Mahalanobis distance metriclearning[J]. IEEE Transactions on Neural Networks,2010,21(9):1524–1530.
[144] WOZNICA A, KALOUSIS A. A new framework for dissimilarity and simi-larity learning[C]//Pacific-Asia Conference on Knowledge Discovery and DataMining. Hyderabad: Springer-Verlag,2010:386–397.
[145] YU J, AMORES J, SEBE N, et al. Distance learning for similarity estima-tion[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(3):451–462.
[146] FUKUNAGA K. Introduction to Statistical Pattern Recognition[M]. New York:Academic Press,1999.
[147] GOLUB G H, LOAN C V. Matrix Computation[M]. Baltimore: The JohnsHopking University Press,1996.
[148] LI L, LI F. What, where, and who? Classifying events by scene and objectrecognition[C]//IEEE Internatioanl Conference on Computer Vision. Rio deJaneiro: IEEE,2007:1–8.
[149] ZHU J, HOI S, LYU M, et al. Near-duplicate keyframe retrieval by nonrigid im-age matching[C]//ACM International Conference on Multimedia. Vancouver:ACM,2008:41–50.