一种基于子空间学习的图像标签推荐方法

doi:10.3969/j.issn.1006-2475.2016.03.014

摘要/Abstract

摘要： 以Flickr和Picasa为代表的社会化图像分享网站允许用户对图像使用标签进行标注，帮助用户更加方便高效地组织和检索图像资源。为了能使用户对所标注的图像提供高质量的标签，自动化图像标签推荐系统成了近年来的热门研究领域。以往的图像标签推荐系统在解决标签推荐冷启动问题上只是简单地利用标签频率信息或者图像的视觉特征相似性进行标签推荐，忽略了图像视觉内容和标签内容之间的关系，往往导致标签推荐结果不是特别理想。本文提出一种新的图像标签推荐方法，该方法利用矩阵分解算法从训练数据集中学习，得到一个图像视觉特征和标签内容语义共享的隐式子空间。对于一幅未打任何标签的新图像，可以利用训练得到的线性转换矩阵将其视觉特征向量映射到隐式子空间中，然后计算得到与各个标签的关联程度进行推荐。本文提出的方法在NUS-WIDE的数据集上进行验证，实验结果比现有代表性方法有大幅提高，表明了该方法的有效性。

关键词: 社会化标签, 社会化图片, 子空间学习, 推荐系统, 冷启动

Abstract: Represented by Flickr and Picasa, online photo albums allow users to tag images, hoping to make it more convenient as well as efficient to organize and retrieve image resources. Recently, automatic tag recommendation system has become a hot research field considering the increasing request that high-quality tags be provided. In this thesis, a new method for tag recommendation system is proposed. Unlike the traditional one which only depends on frequency information or visual feature similarity while neglecting the relation between visual content and the semantic meaning contained in tags thus leading to unsatisfactory recommendations, the new method can find out a latent subspace shared by visual features and tag contents using matrix factorization. As for an untagged image, recommendations can be made when its visual features are projected into the latent subspace and the relevance level it has with others tags is figured out. This new method has been proved efficient after being tested on NUS-WIDE data set with more satisfactory results.

Key words: social tag , social image, subspace learning, recommender system, cold start

中图分类号:

TP391

祁超. 一种基于子空间学习的图像标签推荐方法[J]. 计算机与现代化, doi: 10.3969/j.issn.1006-2475.2016.03.014.

QI Chao. An Approach of Image Tag Recommendation Based on Subspace Learning Model[J]. Computer and Modernization, doi: 10.3969/j.issn.1006-2475.2016.03.014.

参考文献

[1] Ames M, Naaman M. Why we tag: Motivations for annotation in mobile and online media[C]// Proceedings of the 2007 SIGCHI Conference on Human Factors in Computing Systems. 2007:971-980.

[2] Smeulders A W M, Worring M, Santini S, et al. Content-based image retrieval at the end of the early years[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000,22(12):1349-1380.

[3] Datta R, Joshi D, Li Jia, et al. Image retrieval: Ideas, influences, and trends of the new age[J]. ACM Computing Surveys, 2008,40(2): Article 5.

[4] Yavlinsky A, Schofield E, Ruger S. Automated image annotation using global features and robust nonparametric density estimation[C]// Proceedings of the 2005 International Conference on Image and Video Retrieval. 2005:507-517.

[5] Chua T, Tang Jinhui, Hong Richang, et al. NUS-WIDE: A real-world Web image database from National University of Singapore[C]// Proceedings of the 8th ACM International Conference on Image and Video Retrieval. 2009: Article No. 48.

[6] Sun Aixin, Bhowmick S S, Chong Jun-An. Social image tag recommendation by concept matching[C]// Proceedings of the 19th ACM International Conference on Multimedia. 2011:1181-1184.

[7] Lu Yu-Ta, Yu Shoou-I, Chang Tsung-Chieh, et al. A content-based method to enhance tag recommendation[C]// Proceedings of the 21st International Joint Conference on Artificial Intelligence. 2009:2064-2069.

[8] Sigurbjörnsson B, Van Zwol R. Flickr tag recommendation based on collective knowledge[C]// Proceedings of the 17th International Conference on World Wide Web. 2008:327-336.

[9] Li Xirong, Snoek C G M, Worring M. Unsupervised multi-feature tag relevance learning for social image retrieval[C]// Proceedings of the 2010 ACM International Conference on Image and Video Retrieval. 2010:10-17.

[10] Chen Hong-Ming, Chang Ming-Hsiu, Chang Ping-Chieh, et al. SheepDog: Group and tag recommendation for Flickr photos by automatic search-based learning[C]// Proceedings of the 16th ACM International Conference on Multimedia. 2008:737-740.

[11] Liu Dong, Hua Xian-Sheng, Yang Linjun, et al. Tag ranking[C]// Proceedings of the 18th International Conference on World Wide Web. 2009:351-360.

[12] Bottou L. Large-scale machine learning with stochastic gradient descent[C]// Proceedings of COMPSTAT2010. 2010:177-186.

[13] Wang Yong-sheng. Image tag recommendation algorithm using tensor factorization[J]. Journal of Multimedia, 2014,9(3):416-422.

[14] Lindstaedt S, Mörzinger R, Sorschag R, et al. Automatic image annotation using visual content and folksonomies[J]. Multimedia Tools and Applications, 2009,42(1):97-113.

[15] Negoescu R A, Gatica-Perez D. Analyzing Flickr groups[C]// Proceedings of the 2008 International Conference on Content-based Image and Video Retrieval. 2008:417-426.