PARTIALLY TAGGED IMAGE CLUSTERING | |
Yin, Qiyue; Wu, Shu; Wang, Liang | |
2015 | |
会议名称 | International Conference on Image Processing |
会议录名称 | In Proceedings of the International Conference on Image Processing (ICIP), 2015 |
会议日期 | Sep. 27-30 |
会议地点 | Québec |
摘要 | With the exponential growth of tagged images, researchers are resorting to this high semantic tag information to assist the clustering process and promising clustering results have been obtained. However, users may not tag all of their images or some of the images are partially annotated, and this will lead to big performance degradation, which is rarely considered by pervious works. To alleviate this problem, we propose a new framework for image clustering assisted by partially observed tags. Our model enforces the sparse representation obtained through sparse coding and the latent tag representation learned via matrix factorization to be consistent with the partial image-tag observations. Finally, the partitioning of the database is performed using clustering algorithms (e.g., kmeans) on the sparse representation. Extensive experiments on three real world datasets demonstrate that the proposed model performs better than the state-of-the-art methods. |
关键词 | Image Clustering Multi-view Clustering Sparse Coding |
文献类型 | 会议论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/12339 |
专题 | 智能感知与计算研究中心 |
通讯作者 | Wu, Shu |
作者单位 | 中国科学院自动化研究所 |
第一作者单位 | 中国科学院自动化研究所 |
通讯作者单位 | 中国科学院自动化研究所 |
推荐引用方式 GB/T 7714 | Yin, Qiyue,Wu, Shu,Wang, Liang. PARTIALLY TAGGED IMAGE CLUSTERING[C],2015. |
条目包含的文件 | 下载所有文件 | |||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
Partially tagged ima(574KB) | 会议论文 | 开放获取 | CC BY-NC-SA | 浏览 下载 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论