Image semantic annotation of CMRM based on graph learning

LI Ling; SONG Ying-wei; YANG Xiu-hua; CHEN Yi-jie

doi:10.3788/OPE.20162401.0229

您当前的位置：

首页 >

文章列表页 >

Image semantic annotation of CMRM based on graph learning

更新时间：2020-08-13

- Image semantic annotation of CMRM based on graph learning
- Optics and Precision Engineering Vol. 24, Issue 1, Pages: 229-235(2016)
- 作者机构：
  
  1. 吉林大学通信工程学院,吉林长春,中国,130012
  2. 吉林大学网络中心,吉林长春,130012
- 作者简介：
- 基金信息：
- DOI：10.3788/OPE.20162401.0229
  CLC： TP391
- Received：12 November 2015，
  
  Revised：07 December 2015，
  
  Published：25 January 2016
- 稿件说明：
移动端阅览
李玲, 宋莹玮, 杨秀华等. 应用图学习算法的跨媒体相关模型图像语义标注[J]. 光学精密工程, 2016,24(1): 229-235

LI Ling, SONG Ying-wei, YANG Xiu-hua etc. Image semantic annotation of CMRM based on graph learning[J]. Editorial Office of Optics and Precision Engineering, 2016,24(1): 229-235
李玲, 宋莹玮, 杨秀华等. 应用图学习算法的跨媒体相关模型图像语义标注[J]. 光学精密工程, 2016,24(1): 229-235 DOI： 10.3788/OPE.20162401.0229.

LI Ling, SONG Ying-wei, YANG Xiu-hua etc. Image semantic annotation of CMRM based on graph learning[J]. Editorial Office of Optics and Precision Engineering, 2016,24(1): 229-235 DOI： 10.3788/OPE.20162401.0229.

摘要

针对传统跨媒体相关模型(CMRM)只考虑图像的视觉信息与标注词之间的相关性

忽略标注词之间所具有的语义相关性的问题

本文提出了一种新的基于图学习算法的CMRM图像语义标注方法。该方法首先根据运动领域图片训练集中的标注词

建立运动领域本体来标注图像;然后采用传统的CMRM标注算法对训练集图像进行第一次标注

获得基于概率模型的图像标注结果;最后

根据本体概念相似度

利用图学习方法对第一次标注结果进行修正

在每幅图像的概率关系表中选择概率最大的

个关键词作为最终的标注结果

完成第二次标注。实验结果表明

本文提出的模型的查全率和查准率均高于传统的CMRM算法。

Abstract

The traditional Crossmedia Relevance Model(CMRM) is based on the relevance between visual information and annotation words

while ignoring the inter-word semantic relevance. Therefore

a new CMRM image semantic annotation model based on a graph learning was proposed. Firstly

the ontology of a sport field was established to label the images of the sport field according the annotation words in an image training set. Then

the traditional CMRM was adopted in the training images to complete the basic image annotations and obtain the image annotation result based on a probability model. Finally

the graph learning was used to refine the basic image annotations based on ontology concept similarity

and the top

keywords in the probability table for each image were chosen as the final annotation results. Experimental results show that the recall and precision of the proposed model are improved as compared with those of the traditional CMRMs.

关键词

Keywords

references

IM D H, PARK G D. Linked tag: image annotation using semantic relationships between image tags[J]. Multimedia Tools Appl, 2015, 74(7):2273-2287.

吴晓雨,何彦,杨磊,等. 基于改进形状上下文特征的二值图像检索[J]. 光学精密工程,2015,23(1): 302-309. WU X Y, HE Y, YANG L, et al.. Binary image retrieval based on improved shape context algorithm [J]. Opt. Precision Eng., 2015, 23(1): 302-309. (in Chinese)

汪华章,何小海,宰文姣. 基于局部和全局特征融合的图像检索[J]. 光学精密工程, 2008, 16(6):1098-1104. WANG H ZH, HE X H, ZAI W J. Image retrieval based on combining local and global features[J]. Opt. Precision Eng., 2008, 16(6):1098-1104. (in Chinese)

WEI W, GAO G L. Image annotation with nearest neighbor based on semantic information[C]. Proceedings of the 2015 Chinese Intelligent Automation Conference, 2015, 336: 345-352.

ZHANG D S,ISLAM M M, LU G J. A review on automatic image annotation techniques[J]. Pattern Recognition, 2012, 45(1): 346-362.

BLEI D M. Probabilistic topic models[J].Communications of the ACM, 2012, 55(4): 77-84.

刘杰,杜军平. 基于潜在主题融合的跨媒体图像语义标注[J]. 电子学报,2014, 42(5): 987-992. LIU J, DU J P. Latent topic fusion-based cross-media image semantic annotation[J]. Acta Electronica Sinica, 2014, 42(5): 987-992. (in Chinese)

JEON J, LAVRENKO V, MAMMATHA R. Automatic image annotation and retrieval using cross-media relevance model[C]. in Proceedings of the Twenty-Sixth Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, 2003: 119-126.

LAVRENKO V, MANMATHA R, JEON J. A model for learning the semantics of pictures[C]. 17th Annual Conference on Neutral Information Processing system(NIPS), 2003.

SONG H Y, LI X F, WANG P J.Automatic image annotation based on improved relevance model [C]. Asia-Pacific Conference on Information Processing,2009, 2: 59-62.

BANNOUR H, HUDELOT C. Building and using fuzzy multimedia ontologies for semantic image annotation[J]. Multimedia Tools and Application, 2014, 72(3):2107-2141.

ESPINOZA-MOLINA D, NIKOLAOU C, DUMITRU CO, et al..Very-high-resolution SAR images and linked open data analytics based on ontologies[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2015,8(4):1696-1708.

王睿,朱正丹. 融合全局-颜色信息的尺度不变特征变换[J]. 光学精密工程,2015,23(1): 295-301. WANG R, ZHU ZH D. SIFT matching with color invariant characteristics and global context[J]. Opt. Precision Eng., 2015, 23(1): 295-301. (in Chinese)

赵东,赵宏伟,于繁华. 动态多目标优化的运动物体图像分割[J]. 光学精密工程,2015,23(7): 2109-2116. ZHAO D, ZHAO H W, YU F H. Moving object image segmentation by dynamic multi-objective optimization [J]. Opt. Precision Eng., 2015, 23(7): 2109-2116. (in Chinese)

Views

413

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Extraction method for fluorescence information from droplet digital PCR images

Imaging MTF of space camera under vibration and simulation

Improved algorithm for motion image analysis based on optical flow and its application

Related Author

Shanshan LI

Zichao WANG

Chengzhuang YU

Chunyang WEI

Junwei LI

ZHANG Ying

TU Da-wei

JIANG Ji-liang

Related Institution

School of Mechanical Engineering， Hebei University of Technology

Hebei Key Laboratory of Robotic Sensing and Human-Robot Interactions

State Key Laboratory of Reliability and Intelligence of Electrical Equipment

Institute of Biophysics， Hebei University of Technology

Changchun University of Science and Technology

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰