Tag refinement by regularized LDA Export

@inproceedings{Xu2009tag, abstract = {Tagging is nowadays the most prevalent and practical way to make images searchable. However, in reality many tags are irrelevant to image content. To refine the tags, previous solutions usually mine tag relevance relying on the tag similarity estimated right from the corpus to be refined. The calculation of tag similarity is affected by the noisy tags in the corpus, which is not conducive to estimate accurate tag relevance. In this paper, we propose to do tag refinement from the angle of topic modeling. In the proposed scheme, tag similarity and tag relevance are jointly estimated in an iterative manner, so that they can benefit from each other. Specifically, a novel graphical model, regularized Latent Dirichlet Allocation (rLDA), is presented. It facilitates the topic modeling by exploiting both the statistics of tags and visual affinities of images in the corpus. The experiments on tag ranking and image retrieval demonstrate the advantages of the proposed method.}, address = {New York, NY, USA}, author = {Xu, Hao and Wang, Jingdong and Hua, Xian S. and Li, Shipeng}, booktitle = {MM '09: Proceedings of the seventeen ACM international conference on Multimedia}, citeulike-article-id = {6079844}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1631272.1631359}, citeulike-linkout-1 = {http://dx.doi.org/10.1145/1631272.1631359}, doi = {10.1145/1631272.1631359}, isbn = {978-1-60558-608-3}, keywords = {lda, refinement, tag}, location = {Beijing, China}, pages = {573--576}, posted-at = {2009-11-06 03:48:16}, priority = {2}, publisher = {ACM}, title = {Tag refinement by regularized LDA}, url = {http://dx.doi.org/10.1145/1631272.1631359}, year = {2009} }

TY - CONF ID - Xu2009tag L3 - citeulike-article-id:6079844 N2 - Tagging is nowadays the most prevalent and practical way to make images searchable. However, in reality many tags are irrelevant to image content. To refine the tags, previous solutions usually mine tag relevance relying on the tag similarity estimated right from the corpus to be refined. The calculation of tag similarity is affected by the noisy tags in the corpus, which is not conducive to estimate accurate tag relevance. In this paper, we propose to do tag refinement from the angle of topic modeling. In the proposed scheme, tag similarity and tag relevance are jointly estimated in an iterative manner, so that they can benefit from each other. Specifically, a novel graphical model, regularized Latent Dirichlet Allocation (rLDA), is presented. It facilitates the topic modeling by exploiting both the statistics of tags and visual affinities of images in the corpus. The experiments on tag ranking and image retrieval demonstrate the advantages of the proposed method. CY - Beijing, China SN - 978-1-60558-608-3 AD - New York, NY, USA EP - 576 TI - Tag refinement by regularized LDA PB - ACM T2 - MM '09: Proceedings of the seventeen ACM international conference on Multimedia SP - 573 KW - lda KW - refinement KW - tag AU - Xu, Hao AU - Wang, Jingdong AU - Hua, Xian AU - Li, Shipeng PY - 2009/// UR - http://dx.doi.org/10.1145/1631272.1631359 ER -