Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982687
Jun Lei, Guohui Li, Jun Zhang, Dan Lu, Qiang Guo
We propose online learning algorithms for structural SVM that has promising applications in large-scale learning. A framework is introduced for analyzing the online learning of structural SVM from primal perspective to dual perspective. The task of minimizing the primal objective function is converted to incremental increasing of the dual objective function. The model's parameter is learned through updating dual coefficients. We propose two update schemes: all outputs update scheme and most violated output update scheme. The first scheme updates dual coefficients of all the outputs, while the second schemes only updated dual coefficients of the most violated output. The performance of structural SVM is improved in online learning process. Experimental results on multiclass classification task and sequence tagging task show that our online learning algorithms achieve satisfying accuracy while reducing the computational complexity.
{"title":"Online structural SVM learning by dual ascending procedure","authors":"Jun Lei, Guohui Li, Jun Zhang, Dan Lu, Qiang Guo","doi":"10.1109/SPAC.2014.6982687","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982687","url":null,"abstract":"We propose online learning algorithms for structural SVM that has promising applications in large-scale learning. A framework is introduced for analyzing the online learning of structural SVM from primal perspective to dual perspective. The task of minimizing the primal objective function is converted to incremental increasing of the dual objective function. The model's parameter is learned through updating dual coefficients. We propose two update schemes: all outputs update scheme and most violated output update scheme. The first scheme updates dual coefficients of all the outputs, while the second schemes only updated dual coefficients of the most violated output. The performance of structural SVM is improved in online learning process. Experimental results on multiclass classification task and sequence tagging task show that our online learning algorithms achieve satisfying accuracy while reducing the computational complexity.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128951225","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982712
Min Jiang, Xiaozhou Zhou, Shijie Yao, Zhaohui Gan
Extracting non-rigid object from images can be used in object recognition, medical image analysis, video monitoring, etc. In order to improve the efficiency and accuracy of visual object extraction, we design a candidate shape generator based on a mixture strategy, called mixture generator, it combines the image data driven method with model parameter driven method, and tends to generate valid shape in area which has a high shape prior density value by exploiting the GPDM model, so the efficiency of search is greatly improved. To prove the accuracy of our mixture generator, we have done experiments under the framework of global optimization algorithm (simulated annealing) on the FGNET face database. Experiments show that, compared with traditional ASM algorithm, our method is not only insensitive to initialization conditions, but also can put up with clutters and realize a more robust object extraction.
{"title":"A candidate solutions generator based on mixed strategy for non-rigid object extraction","authors":"Min Jiang, Xiaozhou Zhou, Shijie Yao, Zhaohui Gan","doi":"10.1109/SPAC.2014.6982712","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982712","url":null,"abstract":"Extracting non-rigid object from images can be used in object recognition, medical image analysis, video monitoring, etc. In order to improve the efficiency and accuracy of visual object extraction, we design a candidate shape generator based on a mixture strategy, called mixture generator, it combines the image data driven method with model parameter driven method, and tends to generate valid shape in area which has a high shape prior density value by exploiting the GPDM model, so the efficiency of search is greatly improved. To prove the accuracy of our mixture generator, we have done experiments under the framework of global optimization algorithm (simulated annealing) on the FGNET face database. Experiments show that, compared with traditional ASM algorithm, our method is not only insensitive to initialization conditions, but also can put up with clutters and realize a more robust object extraction.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"107 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123236385","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982656
Xin Xu, Weiwei Wu
An adaptive salient region detection method is proposed in this study, which combines LAB and RGB feature space and fused the color and contrast features. This algorithm first extracts the color feature of each image block in the LAB space and the contrast feature in the RGB space, and then fuses the color feature saliency map and the contrast feature saliency map using the principal component analysis (PCA) method which can effectively retain the saliency information of color and contrast, at last, this research extracts the salient region by setting a adaptive threshold. Compared with other detection methods, the proposed method is accurate and highlights the salient region uniformly, the detection results are more in line with the observations of human eyes.
{"title":"An adaptive computational method for color contrast based salient region detection","authors":"Xin Xu, Weiwei Wu","doi":"10.1109/SPAC.2014.6982656","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982656","url":null,"abstract":"An adaptive salient region detection method is proposed in this study, which combines LAB and RGB feature space and fused the color and contrast features. This algorithm first extracts the color feature of each image block in the LAB space and the contrast feature in the RGB space, and then fuses the color feature saliency map and the contrast feature saliency map using the principal component analysis (PCA) method which can effectively retain the saliency information of color and contrast, at last, this research extracts the salient region by setting a adaptive threshold. Compared with other detection methods, the proposed method is accurate and highlights the salient region uniformly, the detection results are more in line with the observations of human eyes.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"141 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131190114","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982726
Wenju Zhang, Naiyang Guan, Xuhui Huang, Zhigang Luo, Jianwu Li
Link partition clusters edges of a complex network to discover its overlapping communities. Due to Its effectiveness, link partition has attracted much attentions from the network science community. However, since link partition assigns each edge of a network to unique community, it cannot detect the disjoint communities. To overcome this deficiency, this paper proposes a link partition on asymmetric weighted graph (LPAWG) method for detecting overlapping communities. Particularly, LPAWG divides each edge into two parts to distinguish the roles of connected nodes. This strategy biases edges to a specific node and helps assigning each node to its affiliated community. Since LPAWG introduces more edges than those in the original network, it cannot efficiently detect communities from some networks with relative large amount of edges. We therefore aggregate the line graph of LPAWG to shrink its scale. Experimental results of community detection on both synthetic datasets and the realworld networks show the effectiveness of LPAWG comparing with the representative methods.
{"title":"Overlapping community detection via link partition of asymmetric weighted graph","authors":"Wenju Zhang, Naiyang Guan, Xuhui Huang, Zhigang Luo, Jianwu Li","doi":"10.1109/SPAC.2014.6982726","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982726","url":null,"abstract":"Link partition clusters edges of a complex network to discover its overlapping communities. Due to Its effectiveness, link partition has attracted much attentions from the network science community. However, since link partition assigns each edge of a network to unique community, it cannot detect the disjoint communities. To overcome this deficiency, this paper proposes a link partition on asymmetric weighted graph (LPAWG) method for detecting overlapping communities. Particularly, LPAWG divides each edge into two parts to distinguish the roles of connected nodes. This strategy biases edges to a specific node and helps assigning each node to its affiliated community. Since LPAWG introduces more edges than those in the original network, it cannot efficiently detect communities from some networks with relative large amount of edges. We therefore aggregate the line graph of LPAWG to shrink its scale. Experimental results of community detection on both synthetic datasets and the realworld networks show the effectiveness of LPAWG comparing with the representative methods.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128067575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982684
Nan Zhu, Xinbo Gao, Cheng Deng
Over the past years, research on digital image forensics has become a hot topic in multimedia security. Among various forensics technologies, image resampling detection has become a standard detection tool in image forensics. Furthermore, examining parameters of geometric transformations such as scaling factors or rotation angles is very useful for exploring an image's overall processing history. In this paper, we propose a novel image scaling factor estimation method based on normalized energy density and learning to rank, which can not only effectively eliminate the long-known ambiguity between upscaling and downscaling in the analysis of resampling but also accurately estimate the factors of weak scaling, i.e., the scaling factors near 1. Empirical experiments on extensive images with different scaling factors demonstrate the effectiveness of our proposed method.
{"title":"Image scaling factor estimation based on normalized energy density and learning to rank","authors":"Nan Zhu, Xinbo Gao, Cheng Deng","doi":"10.1109/SPAC.2014.6982684","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982684","url":null,"abstract":"Over the past years, research on digital image forensics has become a hot topic in multimedia security. Among various forensics technologies, image resampling detection has become a standard detection tool in image forensics. Furthermore, examining parameters of geometric transformations such as scaling factors or rotation angles is very useful for exploring an image's overall processing history. In this paper, we propose a novel image scaling factor estimation method based on normalized energy density and learning to rank, which can not only effectively eliminate the long-known ambiguity between upscaling and downscaling in the analysis of resampling but also accurately estimate the factors of weak scaling, i.e., the scaling factors near 1. Empirical experiments on extensive images with different scaling factors demonstrate the effectiveness of our proposed method.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"247 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115659391","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982680
Weihua Ou, Kesheng Zhang, Xinge You, Fei Long
Real data often are comprised of multiple modalities or different views, which provide complementary and consensus information to each other. Exploring those information is important for the multi-view data clustering and classification. Multiview embedding is an effective method for multiple view data which uncovers the common latent structure shared by different views. Previous studies assumed that each view is clean, or at least there are not contaminated by noises. However, in real tasks, it is often that every view might be suffered from noises or even some views are partially missing, which renders the traditional multi-view embedding algorithm fail to those cases. In this paper, we propose a novel multi-view embedding algorithm via robust joint nonnegative matrix factorization. We utilize the correntropy induced metric to measure the reconstruction error for each view, which are robust to the noises by assigning different weight for different entries. In order to uncover the common subspace shared by different views, we define a consensus matrix subspace to constrain the disagreement of different views. For the non-convex objective function, we formulate it into half quadratic minimization and solve it via update scheme efficiently. The experiments results show its effectiveness and robustness in multiview clustering.
{"title":"Multi-view embedding learning via robust joint nonnegative matrix factorization","authors":"Weihua Ou, Kesheng Zhang, Xinge You, Fei Long","doi":"10.1109/SPAC.2014.6982680","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982680","url":null,"abstract":"Real data often are comprised of multiple modalities or different views, which provide complementary and consensus information to each other. Exploring those information is important for the multi-view data clustering and classification. Multiview embedding is an effective method for multiple view data which uncovers the common latent structure shared by different views. Previous studies assumed that each view is clean, or at least there are not contaminated by noises. However, in real tasks, it is often that every view might be suffered from noises or even some views are partially missing, which renders the traditional multi-view embedding algorithm fail to those cases. In this paper, we propose a novel multi-view embedding algorithm via robust joint nonnegative matrix factorization. We utilize the correntropy induced metric to measure the reconstruction error for each view, which are robust to the noises by assigning different weight for different entries. In order to uncover the common subspace shared by different views, we define a consensus matrix subspace to constrain the disagreement of different views. For the non-convex objective function, we formulate it into half quadratic minimization and solve it via update scheme efficiently. The experiments results show its effectiveness and robustness in multiview clustering.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"235 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126043224","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982701
Yiu-ming Cheung, Junping Deng
Local Binary Pattern (LBP) is a simple yet powerful method for image feature extraction in pattern recognition and image processing. However, the LBP operator of each pixel mainly depends on its neighboring pixels and emphasizes on local information too much. From the practical viewpoint, the information is quite limited if we consider the LBP operator in isolation, especially for a large image. To deal with this issue, we propose ultra LBP (U-LBP), which consider the relationship among different LBP operators. The proposed method cannot only get the local but also ultra local information. The effectiveness of the proposed algorithm is investigated on gender recognition and digit recognition, respectively. The experimental results show that the proposed method outperforms the traditional LBP.
{"title":"Ultra local binary pattern for image texture analysis","authors":"Yiu-ming Cheung, Junping Deng","doi":"10.1109/SPAC.2014.6982701","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982701","url":null,"abstract":"Local Binary Pattern (LBP) is a simple yet powerful method for image feature extraction in pattern recognition and image processing. However, the LBP operator of each pixel mainly depends on its neighboring pixels and emphasizes on local information too much. From the practical viewpoint, the information is quite limited if we consider the LBP operator in isolation, especially for a large image. To deal with this issue, we propose ultra LBP (U-LBP), which consider the relationship among different LBP operators. The proposed method cannot only get the local but also ultra local information. The effectiveness of the proposed algorithm is investigated on gender recognition and digit recognition, respectively. The experimental results show that the proposed method outperforms the traditional LBP.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129377662","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Particle filter and sparse representation have been successfully applied to visual tracking in computer vision community. This paper proposes an adaptive weighted coarse-to-fine sparse tracking(WCFT) method based on particle filter framework. In this method, two series of templates, coarse templates and fine templates, are used to represent two different stages of human vision perception process respectively. Besides, the regularization parameter(weight) of each template is adapted according to its significance in representing the target. We also prove that our problem can be solved using an accelerated proximal gradient(APG) method. Moreover, we prove that the outstanding L1 tracker is a special case of our model and our method is more effective and efficient in general. The superiority of our system over current state-of-art tracking methods is demonstrated by a set of comprehensive experiments on public data sets.
{"title":"Robust weighted coarse-to-fine sparse tracking","authors":"Boxuan Zhong, Zijing Chen, Xinge You, Luoqing Li, Y. Xie, Shujian Yu","doi":"10.1109/SPAC.2014.6982648","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982648","url":null,"abstract":"Particle filter and sparse representation have been successfully applied to visual tracking in computer vision community. This paper proposes an adaptive weighted coarse-to-fine sparse tracking(WCFT) method based on particle filter framework. In this method, two series of templates, coarse templates and fine templates, are used to represent two different stages of human vision perception process respectively. Besides, the regularization parameter(weight) of each template is adapted according to its significance in representing the target. We also prove that our problem can be solved using an accelerated proximal gradient(APG) method. Moreover, we prove that the outstanding L1 tracker is a special case of our model and our method is more effective and efficient in general. The superiority of our system over current state-of-art tracking methods is demonstrated by a set of comprehensive experiments on public data sets.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124565604","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982692
Lianbo Zhang, Dapeng Tao, Weifeng Liu
Image segmentation is a common topic in image processing. Many methods has been used in image segmentation, such as Graph cut, threshold-based. However, these methods can't work with high precision. Among these method, SVM is used as a good tool for classification, as we treat image segmentation as a problem of classification. To solve the problem above and get better segmentation result as well as high precision, we add Laplacian regularization to SVM algorithm to get a new algorithm i.e. Laplacian regularized active learning for image segmentation. Our algorithm considers distance between pixels when segmenting a picture, which is executed by Laplacian regularization. Experiments demonstrate that our algorithm perform better in comparison with common SVM algorithm.
{"title":"Laplacian regularized active learning for image segmentation","authors":"Lianbo Zhang, Dapeng Tao, Weifeng Liu","doi":"10.1109/SPAC.2014.6982692","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982692","url":null,"abstract":"Image segmentation is a common topic in image processing. Many methods has been used in image segmentation, such as Graph cut, threshold-based. However, these methods can't work with high precision. Among these method, SVM is used as a good tool for classification, as we treat image segmentation as a problem of classification. To solve the problem above and get better segmentation result as well as high precision, we add Laplacian regularization to SVM algorithm to get a new algorithm i.e. Laplacian regularized active learning for image segmentation. Our algorithm considers distance between pixels when segmenting a picture, which is executed by Laplacian regularization. Experiments demonstrate that our algorithm perform better in comparison with common SVM algorithm.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"36 8","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131485570","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-12-15DOI: 10.1109/SPAC.2014.6982688
Yang Li, Dapeng Tao, Weifeng Liu, Yanjiang Wang
Semi-supervised learning algorithms that combine labeled and unlabeled data receive significant interests in recent years and are successfully deployed in many practical data mining applications. Manifold regularization, one of the most representative works, tries to explore the geometry of the intrinsic data probability distribution by penalizing the classification function along the implicit manifold. Although existing manifold regularization, including Laplacian regularization (LR) and Hessian regularization (HR), yields significant benefits for partially labeled classification, it is observed that LR suffers from the poor generalization and HR exhibits the characteristic of instability, both manifold regularization could not accurately reflect the ground-truth. To remedy the problems in single manifold regularization and approximate the intrinsic manifold, we propose Manifold Regularized Co-Training (Co-Re) framework, which combines the manifold regularization (LR and HR) and the algorithm co-training. Extensive experiments on the USAA video dataset are conducted and validate the effectiveness of Co-Re by comparing it with baseline manifold regularization algorithms.
{"title":"Co-regularization for classification","authors":"Yang Li, Dapeng Tao, Weifeng Liu, Yanjiang Wang","doi":"10.1109/SPAC.2014.6982688","DOIUrl":"https://doi.org/10.1109/SPAC.2014.6982688","url":null,"abstract":"Semi-supervised learning algorithms that combine labeled and unlabeled data receive significant interests in recent years and are successfully deployed in many practical data mining applications. Manifold regularization, one of the most representative works, tries to explore the geometry of the intrinsic data probability distribution by penalizing the classification function along the implicit manifold. Although existing manifold regularization, including Laplacian regularization (LR) and Hessian regularization (HR), yields significant benefits for partially labeled classification, it is observed that LR suffers from the poor generalization and HR exhibits the characteristic of instability, both manifold regularization could not accurately reflect the ground-truth. To remedy the problems in single manifold regularization and approximate the intrinsic manifold, we propose Manifold Regularized Co-Training (Co-Re) framework, which combines the manifold regularization (LR and HR) and the algorithm co-training. Extensive experiments on the USAA video dataset are conducted and validate the effectiveness of Co-Re by comparing it with baseline manifold regularization algorithms.","PeriodicalId":326246,"journal":{"name":"Proceedings 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131555671","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：