Subspace learning based active learning for image retrieval

	Subspace learning based active learning for image retrieval
	Biao Niu; Yifan Zhang; Jinqiao Wang; Jian Cheng; Hanqing Lu
	2013
会议名称	2013 IEEE International Conference on Multimedia and Expo Workshops (ICMEW)
会议录名称	IEEE International Conference on Multimedia and Expo Workshops (ICMEW)
页码	1-4
会议日期	2013
会议地点	Fairmont San Jose，USA
摘要	The goal of relevance feedback is to improve the performance of image retrieval by leveraging the labeling of human. It is helpful to introduce active learning method in relevance feedback to alleviate the human burden. In the traditional active learning the samples which can improve the classifi- er the most if they were labeled are selected for the user’s labeling. However, the change of the geometrical structure of the data distribution caused by such expensive labeled sam- ples is not fully exploited. By mining user’s labeling infor- mation, we can reduce the original feature space dimension to ease the classifier’s instability brought by the small sam- ple size. In this paper, we propose a novel batch mode ac- tive learning method for informative data selection. The la- beled samples are not only used to retrain the classifier, but to learn a subspace which efficiently encodes user’s inten- tion as well. Especially, a scheme of certainty propagation on the subspace effectively integrates uncertainty sampling and subspace learning into the proposed Subspace learning based batch mode Active Learning method (SubAL) in rele- vance feedback. Extensive experiments on publicly available dataset shows that the proposed method is promising.
关键词	Image Retrieval Subspace Learning
文献类型	会议论文
条目标识符	http://ir.ia.ac.cn/handle/173211/4703
专题	紫东太初大模型研究中心_图像与视频分析
通讯作者	Jinqiao Wang
推荐引用方式 GB/T 7714	Biao Niu,Yifan Zhang,Jinqiao Wang,et al. Subspace learning based active learning for image retrieval[C],2013:1-4.

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
Subspace learning ba（141KB）	会议论文		开放获取	CC BY-NC-SA	浏览下载