王碧莹,
韩力英,
周亚同
河北工业大学电子信息工程学院 天津 300401
基金项目:教育部春晖计划项目(Z2017015)
详细信息
作者简介:唐红梅：女，1968年生，副教授，研究方向为数字图像处理、模式识别
王碧莹：女，1993年生，硕士生，研究方向为数字图像处理、模式识别
韩力英：女，1977年生，讲师，研究方向为图像处理、机器学习
周亚同：男，1973年生，教授，研究方向为机器学习、模式识别
通讯作者:唐红梅 hmtang2005@163.com
中图分类号:TP391.41计量
文章访问数:1590
HTML全文浏览量:874
PDF下载量:60
被引次数:0
出版历程
收稿日期:2019-02-21
修回日期:2019-05-28
网络出版日期:2019-06-04
刊出日期:2019-10-01
Image Saliency Detection Based on Object Compactness and Regional Homogeneity Strategy
Hongmei TANG,,Biying WANG,
Liying HAN,
Yatong ZHOU
School of Electronics and Information Engineering, Hebei University of Technology, Tianjin 300401, China
Funds:Chunhui project of the Ministry of Education (Z2017015)
摘要
摘要:针对基于图模型的显著性检测算法中节点间特征差异描述不准确的问题，该文提出一种目标紧密性与区域同质性策略相结合的图像显著性检测算法。区别于常用的图模型，该算法建立更贴近人眼视觉系统的稀疏图结构与新颖的区域同质性图结构，以便描述图像前景内部的关联性与前景背景间的差异性，从而摒弃众多节点的冗余连接，强化节点局部空间关系；并且结合聚类簇紧密性采取流形排序的方式形成显著图，利用背景区域簇的相似性，引入背景置信度进行显著性优化，最终得到精细的检测结果。在4个基准数据集上与4种基于图模型的流行算法对比，该算法能清晰地突出显著区域，且在多种综合指标评估中，具备更优越的性能。
关键词:图模型/
目标紧密性/
区域同质性/
流形排序/
显著性检测
Abstract:Considering the inaccurate description of feature differences between nodes in the graph-based saliency detection algorithm, an image saliency detection algorithm combining object compactness and regional homogeneity strategy is proposed. Different from the commonly used graph-based model, a sparse graph-based structure closer to the human visual system and a novel regional homogeneity graph-based structure are established. They are used to describe the correlation within the foreground and the difference between foreground and background. Therefore, many redundant connections of nodes are eliminated and the local spatial relationship of nodes is strengthened. Then the clusters are combined to form a saliency map by means of manifold ranking. Finally, the background confidence is introduced for saliency optimization by the similarity of the background region clusters and the final detection result is obtained. Compared with 4 popular graph-based algorithms on the four benchmark datasets, the proposed algorithm can highlight the salient regions clearly and has better performance in the evaluation of multiple comprehensive indicators.
Key words:Graph-based model/
Object compactness/
Regional homogeneity/
Manifold ranking/
Saliency detection
PDF全文下载地址:
https://jeit.ac.cn/article/exportPdf?id=0e584c53-da0e-461d-88d4-931b1cc4b871
