基于语境相关图传播的图像标注改善方法

doi:10.6040/j.issn.1672-3961.1.2015.316

山东大学学报(工学版) ›› 2016, Vol. 46 ›› Issue (5): 1-6.doi: 10.6040/j.issn.1672-3961.1.2015.316

• • 下一篇

基于语境相关图传播的图像标注改善方法

田枫¹,刘卓炫¹,尚福华^1*,沈旭昆²,王梅¹,王浩畅¹

1. 东北石油大学计算机与信息技术学院, 黑龙江大庆 163318;2. 北京航空航天大学虚拟现实技术与系统国家重点试验室, 北京 100191

收稿日期:2015-03-01 出版日期:2016-10-20 发布日期:2015-03-01
通讯作者: 尚福华(1962— ),男,吉林延吉人,教授,博士,主要研究方向为机器学习.E-mail:shangfh@126.com E-mail:tianfeng1980@163.com
作者简介:田枫(1980— ),男,黑龙江安达人,副教授,博士,主要研究方向为多媒体数据理解.E-mail:tianfeng1980@163.com
基金资助:
国家自然科学基金资助项目(61502094,61402099);黑龙江省自然科学基金资助项目(F2016002,F2015020);黑龙江省教育科学规划重点课题资助项目(GJB1215019)

Image annotation refinement based on contextual graph diffusion

TIAN Feng¹, LIU Zhuoxuan¹, SHANG Fuhua^1*, SHEN Xukun², WANG Mei¹, WANG Haochang¹

1. School of Computer and Information Technology, Northeast Petroleum University, Daqing 163318, Heilongjiang, China;
2. State Key Laboratory of Virtual Reality Technology and Systems, BeiHang University, Beijing 100191, China

Received:2015-03-01 Online:2016-10-20 Published:2015-03-01

摘要/Abstract

摘要： 提出一种图像标注改善方法,利用数据集蕴含的语境相关信息进行标注改善。构建标签相关图和视觉内容相关图,利用正则化框架将标注改善问题描述为两个无向加权图上的损失函数最小化问题。采用数据分割,逐次优化和放松约束的策略,获得该问题的近似解。该方法充分利用标签的语境相关信息和图像内容相关信息,对数据集分割的粒度具有较好的鲁棒性,具备近似线性的时间复杂度。测试结果表明,该方法适用于大规模数据集,性能优于其它对比方法,可以较大幅度的提升图像标注性能。

关键词: 图像标注, 标注改善, 语境相关图, 语境信息传播, 大规模数据集

Abstract: A new image annotation refinement method was proposed. The initial labels were firstly obtained by annotation methods. Then label relevant graph and visual relevant graph were constructed and mutually reinforced. The semantic optimization problem was formulated into a regularized framework on above undirected weighted graphs. With strategies like data partitioning, successive optimization and constraint relaxation, an approximate optimized solution was got. The refined result could be more related to the content of images by incorporating both visual content and contextual information. Moreover, the proposed method was robust with the partition granularity, and the complexity was approximately linear. Experimental results on large scale web image dataset showed that the proposed method outperformed others, and could achieve both efficiency and capability.

Key words: contextual information diffusion, large scale data set, contextual graph, image annotation, annotation refinement

中图分类号:

TP391

田枫,刘卓炫,尚福华,沈旭昆,王梅,王浩畅. 基于语境相关图传播的图像标注改善方法[J]. 山东大学学报(工学版), 2016, 46(5): 1-6.

TIAN Feng, LIU Zhuoxuan, SHANG Fuhua, SHEN Xukun, WANG Mei, WANG Haochang. Image annotation refinement based on contextual graph diffusion[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2016, 46(5): 1-6.

参考文献

[1] JIN Y H, KHAN L, WANG L. Image annotations by combining multiple evidence & Wordnet[C] //13th Annual ACM International Conference on Multimedia. Singapore: ACM Press, 2005:706-715.
[2] JIN Y H, KHAN L, PRABHAKARAN B. Knowledge based image annotation refinement[J]. Journal of Signal Processing Systems, 2010, 58(3):387-406.
[3] WANG C H, FENG J, ZHANG L, et al. Image annotation refinement using random walk with restarts[C] //14th ACM International Conference on Multimedia. Santa Barbara, USA: ACM Press, 2006:647-650.
[4] WANG C H, FENG J, ZHANG L, et al. Content-based image annotation refinement[C] //IEEE Conference on Computer Vision and Pattern Recognition. Minneapolis, USA: IEEE Press, 2007:1-8.
[5] LI X R, SNOEK C G M, WORRING M. Learning social tag relevance by neighbor voting[J]. IEEE Transactions on Multimedia, 2009:1310-1322
[6] LIU D, HUA X C, YANG L J, et al. Tag ranking[C] //18th Proceedings of the Intenational World Wide Web Conference. Madrid, Spain: ACM Press, 2009:351-360.
[7] CHEN L, XU D, TSANG I W. Tag-based web photo retrieval improved by batch mode re-tagging[C] //23th IEEE Conference on Computer Vision and Pattern Recognition. San Francisco, USA: IEEE Press, 2010:3440-3446.
[8] FAN J P, SHEN Y, ZHOU N. Harvesting large-scale weakly-tagged image databases from the web[C] //23th IEEE Conference on Computer Vision and Pattern Recognition. San Francisco, USA: IEEE Press, 2010:802-809.
[9] RABINOVICH A, VEDALDI A, GALLEGUILLOS C, et al. Objects in context[C] //11th IEEE International Conference on Computer Vision. Rio de Janeiro,Brazil: Institute of Electrical and Electronics Engineers, 2007:1-8.
[10] MALISIEWICZ T, EFROS A A. Beyond categories: the visual memex model for reasoning about object relationships[C] //23th Conference on Advances in Neural Information Processing Systems. Vancouver, Canada: Curran Associates, 2009:1222-1230.
[11] CHOI M J, LIM J J, TORRALBA A. Exploiting hierarchical context on a large database of object categories[C] //23th IEEE Conference on Computer Vision and Pattern Recognition. San Francisco, USA: IEEE Press, 2010:129-136.
[12] LEE Y J, GRAUMAN K. Object-graphs for context-aware category discovery[C] //23th IEEE Conference on Computer Vision and Pattern Recognition. San Francisco, USA: IEEE Press, 2010:1-8.
[13] WANG H, HUANG H, DING C. Image annotation using multi-label correlated Greens function[C] //12th IEEE International Conference on Computer Vision. Kyoto, Japan: IEEE Press, 2009:1-8.
[14] WANG H, HUANG H, DING C. Multi-label feature transform for image classifications[C] //11th European Conference on Computer Vision. Crete, Greece: Springer Press, 2010:793-806.
[15] WANG H, HUANG H, DING C. Image annotation using bi-relational graph of images and semantic labels[C] //24th IEEE Conference on Computer Vision and Pattern Recognition. Boston, USA: IEEE Press, 2011:793-800.
[16] JIANG Y G, QI D, WANG J, et al. Fast semantic diffusion for large-scale context-based image and video annotation[J]. IEEE Transactions on Image Processing, 2012, 21(6):3080-3091.
[17] HUANG S J, ZHOU Z H. Multi-label learning by exploiting label correlations locally[C] //26th National Conference on Artificial Intelligence. Toronto, Canada: AI Access Foundation, 2012:949-955.
[18] YANG Y, WU F, NIE F P. Web and personal image annotation by mining label correlation with relaxed visual graph embedding[J]. IEEE Transactions on Image Processing, 2012, 21(3): 1339-1351.
[19] XIANG Y, ZHOU X D, LIU Z T, et al. Semantic context modeling with maximal margin conditional random fields for automatic image annotation[C] //28th IEEE Conference on Computer Vision and Pattern Recognition. San Francisco, USA: IEEE Press, 2010:3368-3375.
[20] CHUA T S, TANG J H, HONG R H. NUS-WIDE: a real-world web image database from national university of singapore[C] //8th ACM Conference on Image and Video Retrieval. Santorini Island, Greece: ACM Press, 2009:1-9.
[21] HUISKES M J, LEW M S. The MIR flickr retrieval evaluation[C] //1st ACM International Conference on Multimedia Information Retrieval. Vancouver, Canada: ACM, 2008:39-43.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于语境相关图传播的图像标注改善方法

Image annotation refinement based on contextual graph diffusion

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

多维度评价

本文评价

推荐阅读 10

[1]	熊冰妍, 王国胤, 邓维斌. 分级式代价敏感决策树及其在手机换机预测中的应用[J]. 山东大学学报(工学版), 0, (): 36-42.
[2]	王晓初, 王士同, 包芳. 基于概率密度分布一致约束的最小最大概率机图像分类算法[J]. 山东大学学报(工学版), 0, (): 13-21.
[3]	沈冬冬,周风余,栗梦媛,王淑倩,郭仁和. 基于集成深度神经网络的室内无线定位[J]. 山东大学学报 (工学版), 2018, 48(5): 95-102.
[4]	胡建平,李鑫,谢琪,李玲,张道畅. 基于Delaunay三角化的二维无约束优化EMD方法[J]. 山东大学学报 (工学版), 2018, 48(5): 9-15, 37.
[5]	吴晨谋,方志军,黄正能. 基于单目摄像头的主动式驾驶行为分析算法[J]. 山东大学学报 (工学版), 2018, 48(5): 69-76.
[6]	王国新,陈凤东,刘国栋. 基于彩色伪随机编码结构光特征提取方法[J]. 山东大学学报 (工学版), 2018, 48(5): 55-60.
[7]	张璞,刘畅,王永. 基于特征融合和集成学习的建议语句分类模型[J]. 山东大学学报 (工学版), 2018, 48(5): 47-54.
[8]	李广丽,刘斌,朱涛,殷依,张红斌. 基于优选典型相关分量的跨媒体检索模型[J]. 山东大学学报 (工学版), 2018, 48(5): 38-46.
[9]	牟廉明. 自适应特征选择加权k子凸包分类[J]. 山东大学学报 (工学版), 2018, 48(5): 32-37.
[10]	陈海永,余力,刘辉,杨佳博,胡启迪. 基于经验小波的太阳能电池缺陷图像融合[J]. 山东大学学报 (工学版), 2018, 48(5): 24-31.
[11]	张东波,寇涛,许海霞. 基于LDB描述子和局部空间结构匹配的快速场景辨识[J]. 山东大学学报 (工学版), 2018, 48(5): 16-23.
[12]	黄劲潮. 基于快速区域建议网络的图像多目标分割算法[J]. 山东大学学报(工学版), 2018, 48(4): 20-26.
[13]	张宪红,张春蕊. 基于六维前馈神经网络模型的图像增强算法[J]. 山东大学学报(工学版), 2018, 48(4): 10-19.
[14]	江珊珊,杨静,范丽亚. 基于PDEs的图像特征提取方法[J]. 山东大学学报(工学版), 2018, 48(4): 27-36.
[15]	窦婷婷,姚元玺,陈鹏,芦灯. 基于ATP-EMTP的电弧建模及工程仿真[J]. 山东大学学报(工学版), 2018, 48(4): 102-108.