[1]路 晶,金奕江,马少平,等.使用基于SVM的否定概率和法的图像标注[J].智能系统学报,2006,1(01):62-66.
 LU Jing,JIN Yi-jiang,MA Shao-ping,et al.Image annotation using the summation of negative probability based on SVM[J].CAAI Transactions on Intelligent Systems,2006,1(01):62-66.
点击复制

使用基于SVM的否定概率和法的图像标注(/HTML)
分享到:

《智能系统学报》[ISSN:1673-4785/CN:23-1538/TP]

卷:
第1卷
期数:
2006年01期
页码:
62-66
栏目:
出版日期:
2006-03-25

文章信息/Info

Title:
Image annotation using the summation of negative probability based on SVM
文章编号:
1673-4785(2006)01-0062-05
作者:
路 晶金奕江马少平茹立云
清华大学计算机科学与技术系,智能技术与系统国家重点实验室,北京100084
Author(s):
LU Jing JIN Yi-jiang MA Shao-ping RU Li-yun
The State Key Lab of Intelligence Technology System, Department of Computer Science and Technology, Tsinghua University, Beijing 100084,China
关键词:
语义标签否定概率和法成对耦合标注向量
Keywords:
semantic label the summation of negative probability pairwise coupling label-vector
分类号:
TP3
文献标志码:
A
摘要:
在基于内容的图像检索中,建立图像底层视觉特征与高层语义的联系是个难题.对此提出了一种为图像提供语义标签的标注方法.先建立小规模图像库为训练集,库中每个图像标有单一的语义标签,再利用其底层特征,以SVM为子分类器,“否定概率和”法为合成方法构建基于成对耦合方式(PWC)的多类分类器,并对未标注的图像进行分类,结果以N维标注向量表示,实验表明,与一对多方式(OPC)的多类分类器及使用概率和法的PWC相比,“否定概率和”法性能更好.
Abstract:
In the approach of content-based image retrieval, a critical point is to provide maximum support in bridging the semantic gap betwee n lowlevel visual features and highlevel concepts. An annotation procedure f or providing images with semantic labels was proposed. The annotation procedure started with labeling a small set of training images, each with one semantic lab el. An ensemble of support vector machines (SVMs) based on the summation of nega tive probability, which was constructed by pairwise coupling (PWC), was then got with the contentbased image features. It was applied to give the unlabeled im age an N dimension labelvector, thus providing users with a conceptualiz ed ann otation. The ensemble is better than the one per class (OPC) scheme and the PWC based on the summation of probability.

参考文献/References:

[1]SMEULDERS A W M,WORRING M,SANTINI S,et al.Contentbased image ret rie val at the end of the early years[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2000, 22(12):1349-1380.
[2]WANG J, LI J, WIEDERHOLD G. Simplicity: semantics-sensitive integrated matc hing for picture libraries [J]. IEEE Transactions on Pattern Analysis Machine Intelligence, 2001,23(9):947-963.
[3]〖JP3〗SRIHARI R K, ZHANG Z F. A semiautomated image annotation system[J]. IEEE Multimedia, 2000, 7(3): 61-71.
[4]TANG H L,HANKA R,HORACE H S.Histological image retrieval based on sem antic content analysis[J]. IEEE Transactions on Information Technology in Biomedicin e,2003,7(1):26-36.
[5]VAILAYA A, FIGUEIREDO M, JAIN A, et al. A Bayesian framework for semantic c lassification of outdoor vacation images[A]. In Proceedings of SPIE: Storage a nd Retrieval for Image and Video Databases VII[C]. San Jose: USA, 1999.
[6]SYCHAY G, CHANG E, GOH K. Effective image annotation via active learnin g [A]. In Proc IEEE Int Conf Multimedia [C]. Switzerland,2002.
[7]PLATT J. Probabilistic outputs for SVMS and comparisons to regularized like lihood methods[A].In Advances in Large Margin Classifiers[C]. Cambrige, 2000.
[8]茹立云,彭    潇,苏    中,等. 基于内容图像检索中的特征性能评价[J]. 计算机研究与发展,2003,40(11):1566-1570.
RU Liyun, PENG Xiao, SU Zhong, et al. Feature performance evaluation in content -based ima ge retrieval[J]. Journal of Computer Research and Development, 2003, 40(11): 1 566-1570.
[9]GOH K, CHANG E,CHENG K T. SVM binary classifier ensembles for im age classification[A]. In Proc ACM CIKM 2001[C]. New York, USA,2001.
[10]WANG J Z. Corel image galley[OL]. http://wang.ist.psu.edu.2000.

备注/Memo

备注/Memo:
收稿日期:2006-03-02.
基金项目:国家重点基础研究基金资助项目(2004CB31810 8);国家自然科学基金资助项目(60223004,60321002,60503064,60303005);教育部科学技术研究重点项目(104236).
作者简介:
路    晶,1980年生,清华大学计算机系博士研究生,主要研究方向为图像检索、模式识别等.发表论文3篇.
金奕江,1970年生,助研,博士,毕业于清华大学计算机系,主要研究方向为模式识别、智能信息检索等.发表论文10余篇,1998年曾获国家教委科技进步二等奖.
马少平,教授,博士生导师,毕业于清华大学计算机系,主要研究方向为智能信息检索、模式识别等.发表论文50余篇.出版专著3部.承担多项国家自然科学基金,“863”“973”和国际合作项目.1998年获国家教委科技进步二等奖.
更新日期/Last Update: 2009-04-07