基于最大间隔的半监督图像搜索重排序方法
|
摘要
提出一种基于最大间隔原理的半监督图像搜索重排序学习算法。所提算法在最大间隔原理框架下,首先利用超图正则化保持标注及未标注样本在原始空间中的局部近邻关系,增强算法的稳健性;其次,利用少量的标注样本构造优先关系对,将样本间先验的相关性等级信息引入目标函数中以更好地指导重排序模型的学习。在公开数据集MSRA-MM1.0上的实验结果表明所提方法能更好地将符合用户需求的结果靠前优先呈现给用户,提高搜索的准确性。
We propose a max margin based semi-supervised reranking method for multimedia information retrieval.We use hypergraph regularization to preserve the neighborhood of the sample in the original space and introduce the labeled and unlabeled sample information to construct the objective function,so as to achieve full and efficient use of data information for ranking.By using a small amount of annotation samples to construct the priority relationship pairs,the priority information between samples is introduced into the objective function to construct a ranking learning model.This method can show users in priority the results that meet their demand better,and improve the retrieval accuracy.The experimental results on MSRA-MM 1.0 dataset suggest the proposed method provides superior performance compared with several state-of-the-art methods.
We propose a max margin based semi-supervised reranking method for multimedia information retrieval.We use hypergraph regularization to preserve the neighborhood of the sample in the original space and introduce the labeled and unlabeled sample information to construct the objective function,so as to achieve full and efficient use of data information for ranking.By using a small amount of annotation samples to construct the priority relationship pairs,the priority information between samples is introduced into the objective function to construct a ranking learning model.This method can show users in priority the results that meet their demand better,and improve the retrieval accuracy.The experimental results on MSRA-MM 1.0 dataset suggest the proposed method provides superior performance compared with several state-of-the-art methods.
引文
[1]Hong C Q,Zhu J K.Hypergraph-based multiexample ranking with sparse representation for transductive learning image retrieval[J].Neurocomputing,2013,101:94-103.
[2]Xie H,Lu Y M.Content-based image re-ranking technology in search engine[J].Journal of Computer Applications,2013,33(2):460-462.谢辉,陆月明.搜索引擎中基于内容的图像重排序[J].计算机应用,2013,33(2):460-462.
[3]Krapac J,Allan M,Verbeek J,et al.Improving web image search results using query-relative classifier[C]∥IEEE Conference on Computer Vision and Pattern Recognition,2010:1094-1101.
[4]Tu S Q,Xue Y J,Liang Y,et al.Review on RGB-Dimage classification[J].Laser&Optoelectronics Progress,2016,53(6):060003.涂淑琴,薛月菊,梁云,等.RGB-D图像分类方法研究综述[J].激光与光电子学进展,2016,53(6):060003.
[5]Ben-Haim N,Babenko B,Belongie S.Improving web-based image search via content based clustering[C]∥IEEE Conference on Computer Vision and Pattern Recognition Workshop,2006:106.
[6]Zeng T Y,Du F.Image super-resolution reconstruction based on hierarchical clustering[J].Acta Optica Sinica,2018,38(4):0410004.曾台英,杜菲.基于层次聚类的图像超分辨率重建[J].光学学报,2018,38(4):0410004.
[7]Yang L,Hanjalic A.Supervised reranking for web image search[C]∥ACM International Conference on Multimedia,2010:183-192.
[8]Hsu W H,Kennedy L S,Chang S.Video search reranking through random walk over document level context graph[C]∥ACM International Conference on Multimedia,2007:971-980.
[9]Wang X,Liu K,Tang X.Query-specific visual semantic spaces for web image re-ranking[C]∥IEEEConference on Computer Vision and Pattern Recognition,2011:857-864.
[10]Chi M,Zhang P,Zhao Y,et al.Web image retrieval reranking with multi-view clustering[C]∥ACMInternational Conference on World Wide Web,2009:1189-1190.
[11]Liu Y,Mei T.Optimizing visual search reranking via pairwise learning[J].IEEE Transactions on Multimedia,2011,13(2):280-291.
[12]Pang S,Xue J,Gao Z,et al.Image reranking with an alternating optimization[C]∥ACM International Conference on Multimedia,2014:1141-1144.
[13]Page L.The PageRank citation ranking:bringing order to the web[R/OL].(1998-01-29)[2018-01-05].http:∥ilpubs.stanford.edu:8090/422/1/1999-66.pdf.
[14]Pang Y W,Ji Z,Jing P G,et al.Ranking graph embedding for learning to rerank[J].IEEETransactions on Neural Networks and Learning Systems,2013,24(8):1292-1303.
[15]Liang J Y,Gao J W,Chang Y.The research and advances on semi-supervised learning[J].Journal of Shanxi University(Natural Science Edition),2009,32(4):528-534.梁吉业,高嘉伟,常瑜.半监督学习研究进展[J].山西大学学报(自然科学版),2009,32(4):528-534.
[16]Jin Z F F,Hou Z Q,Yu W S,et al.Multiple feature fusion based on covariance matrix for visual tracking[J].Acta Optica Sinica,2017,37(9):0915005.金泽芬芬,侯志强,余旺盛,等.基于协方差矩阵的多特征融合跟踪算法[J].光学学报,2017,37(9):0915005.
[17]Ji Z,Pang Y,He Y,et al.Semi-supervised LPPalgorithms for learning-to-rank-based visual search reranking[J].Information Sciences,2015,302(C):83-93.
[18]Wang L,Shuai J M.Query dependent visual similarity in image search reranking[J].Computer Systems&Application,2010,19(11):66-70.王黎,帅建梅.图像重排序中与查询相关的图像相似性度量[J].计算机系统应用,2010,19(11):66-70.
[19]Rvelin K,Kek J.Cumulated gain-based evaluation of IR techniques[J].ACM Transactions on Information Systems,2002,20(4):422-446.
[20]Herbrich R,Graepel,T,Obermayer K.Large margin rank boundaries for ordinal regression[C]∥Workshop on Advances in Large Margin Classifiers,2000:115-132.
[21]Zan B F,Kong J,Jiang M.Human action recognition based on discriminative collaborative representation classifier[J].Laser&Optoelectronics Progress,2018,55(1):011010.昝宝锋,孔军,蒋敏.基于判别协作表征分类器的人体行为识别[J].激光与光电子学进展,2018,55(1):011010.
[22]He X,Yan S,Hu Y,et al.Face recognition using Laplacian faces[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2005,27(3):328-340.
[23]Zhou B,He Y Q,Wang J.Face recognition based on adaptive neighborhood locality preserving projection algorithm[J].Laser&Optoelectronics Progress,2018,55(3):031010.周博,何宇清,王建.基于自适应近邻局部保持投影算法的人脸识别[J].激光与光电子学进展,2018,55(3):031010.
[2]Xie H,Lu Y M.Content-based image re-ranking technology in search engine[J].Journal of Computer Applications,2013,33(2):460-462.谢辉,陆月明.搜索引擎中基于内容的图像重排序[J].计算机应用,2013,33(2):460-462.
[3]Krapac J,Allan M,Verbeek J,et al.Improving web image search results using query-relative classifier[C]∥IEEE Conference on Computer Vision and Pattern Recognition,2010:1094-1101.
[4]Tu S Q,Xue Y J,Liang Y,et al.Review on RGB-Dimage classification[J].Laser&Optoelectronics Progress,2016,53(6):060003.涂淑琴,薛月菊,梁云,等.RGB-D图像分类方法研究综述[J].激光与光电子学进展,2016,53(6):060003.
[5]Ben-Haim N,Babenko B,Belongie S.Improving web-based image search via content based clustering[C]∥IEEE Conference on Computer Vision and Pattern Recognition Workshop,2006:106.
[6]Zeng T Y,Du F.Image super-resolution reconstruction based on hierarchical clustering[J].Acta Optica Sinica,2018,38(4):0410004.曾台英,杜菲.基于层次聚类的图像超分辨率重建[J].光学学报,2018,38(4):0410004.
[7]Yang L,Hanjalic A.Supervised reranking for web image search[C]∥ACM International Conference on Multimedia,2010:183-192.
[8]Hsu W H,Kennedy L S,Chang S.Video search reranking through random walk over document level context graph[C]∥ACM International Conference on Multimedia,2007:971-980.
[9]Wang X,Liu K,Tang X.Query-specific visual semantic spaces for web image re-ranking[C]∥IEEEConference on Computer Vision and Pattern Recognition,2011:857-864.
[10]Chi M,Zhang P,Zhao Y,et al.Web image retrieval reranking with multi-view clustering[C]∥ACMInternational Conference on World Wide Web,2009:1189-1190.
[11]Liu Y,Mei T.Optimizing visual search reranking via pairwise learning[J].IEEE Transactions on Multimedia,2011,13(2):280-291.
[12]Pang S,Xue J,Gao Z,et al.Image reranking with an alternating optimization[C]∥ACM International Conference on Multimedia,2014:1141-1144.
[13]Page L.The PageRank citation ranking:bringing order to the web[R/OL].(1998-01-29)[2018-01-05].http:∥ilpubs.stanford.edu:8090/422/1/1999-66.pdf.
[14]Pang Y W,Ji Z,Jing P G,et al.Ranking graph embedding for learning to rerank[J].IEEETransactions on Neural Networks and Learning Systems,2013,24(8):1292-1303.
[15]Liang J Y,Gao J W,Chang Y.The research and advances on semi-supervised learning[J].Journal of Shanxi University(Natural Science Edition),2009,32(4):528-534.梁吉业,高嘉伟,常瑜.半监督学习研究进展[J].山西大学学报(自然科学版),2009,32(4):528-534.
[16]Jin Z F F,Hou Z Q,Yu W S,et al.Multiple feature fusion based on covariance matrix for visual tracking[J].Acta Optica Sinica,2017,37(9):0915005.金泽芬芬,侯志强,余旺盛,等.基于协方差矩阵的多特征融合跟踪算法[J].光学学报,2017,37(9):0915005.
[17]Ji Z,Pang Y,He Y,et al.Semi-supervised LPPalgorithms for learning-to-rank-based visual search reranking[J].Information Sciences,2015,302(C):83-93.
[18]Wang L,Shuai J M.Query dependent visual similarity in image search reranking[J].Computer Systems&Application,2010,19(11):66-70.王黎,帅建梅.图像重排序中与查询相关的图像相似性度量[J].计算机系统应用,2010,19(11):66-70.
[19]Rvelin K,Kek J.Cumulated gain-based evaluation of IR techniques[J].ACM Transactions on Information Systems,2002,20(4):422-446.
[20]Herbrich R,Graepel,T,Obermayer K.Large margin rank boundaries for ordinal regression[C]∥Workshop on Advances in Large Margin Classifiers,2000:115-132.
[21]Zan B F,Kong J,Jiang M.Human action recognition based on discriminative collaborative representation classifier[J].Laser&Optoelectronics Progress,2018,55(1):011010.昝宝锋,孔军,蒋敏.基于判别协作表征分类器的人体行为识别[J].激光与光电子学进展,2018,55(1):011010.
[22]He X,Yan S,Hu Y,et al.Face recognition using Laplacian faces[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2005,27(3):328-340.
[23]Zhou B,He Y Q,Wang J.Face recognition based on adaptive neighborhood locality preserving projection algorithm[J].Laser&Optoelectronics Progress,2018,55(3):031010.周博,何宇清,王建.基于自适应近邻局部保持投影算法的人脸识别[J].激光与光电子学进展,2018,55(3):031010.