基于聚类分析的鱼眼图像显著性信息检测方法  

Fish-eye lens image saliency detection based on cluster analysis

作　　者：朱均超[1] 张力丹[1] 冯为嘉[1] 刘畅[1] ZHU Jun-chao ZHANG Li-dan FENG Wei-jia LIU Chang(Tianjin Key Laboratory for Control Theory and Applications in Complicated Systems,College of Automation, Tianjin University of Technology, Tianjin 300384, Chin)

机构地区：[1]天津理工大学自动化学院天津市复杂工业系统控制理论及应用重点实验室,天津300384

出　　处：《计算机工程与设计》2017年第5期1292-1295,1313,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(61172185);天津市高等学校科技发展基金项目(20100705)

摘　　要：为实现对大视角鱼眼图像中显著性信息的快速准确检测,提出一种基于聚类分析的鱼眼图像显著性信息检测算法。由Gabor滤波器获取图像的纹理特征,利用K-means聚类算法依据鱼眼图像内部结构关联对底层特征进行聚类,通过对图像颜色、纹理等底层特征信息的分析,得到符合人眼视觉特点的显著性鱼眼图像,利用F-Measue指标进行结果可靠性检测,验证该方法的可行性和有效性。To detect saliency information in fish-eye lens image clearly, a saliency detection based on cluster analysis method was put forward. Gabor filter was adopted to get the texture feature and K-means algorithm was used to cluster low-level features in fish-eye lens image based on its internal structure. By analyzing those features, the saliency map corresponding to human visual characteristics was found. F-Measure criterion was used to verify the feasibility of the proposed method.

关 键 词：显著性检测 鱼眼图像 聚类分析 GABOR滤波器 K均值算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]