基于显著性区域的数字媒体图像色域边界提取方法  

Color Gamut Boundary Extraction Method for Digital Media Images based on Salient Region

作　　者：吴云轩 WU Yunxuan(School of Cultural Communication,Liming Vocational University,Quanzhou 362000,China;College of Culture and Arts,Graduate University of Mongolia,Ulaanbaatar 14192,Mongolia)

机构地区：[1]黎明职业大学文化传播学院,福建泉州362000 [2]蒙古研究大学文化艺术学院,蒙古国乌兰巴托14192

出　　处：《成都工业学院学报》2024年第3期55-60,共6页Journal of Chengdu Technological University

基　　金：福建省职业教育研究课题(TD2022036)。

摘　　要：为提升色域边界提取的完整性,提出一种基于显著性区域的数字媒体图像色域边界提取方法。通过简单线性迭代聚类算法,超像素分割原始数字媒体图像;利用基于纹理特征的K均值(K-means)聚类方法,修正超像素分割结果中的欠分割区域;利用颜色直方图全局对比算法,计算修正后超像素层的颜色显著性值,结合贝叶斯框架,提取数字媒体图像的显著性区域;以色相角聚类的方式,在显著性区域内,提取数字媒体图像色域边界。实验证明:该方法可有效超像素分割数字媒体图像,得到显著性区域;可有效提取数字媒体图像色域边界;提取的色域边界点空间分布较均匀,即色域边界提取的完整性较优。In order to improve the integrity of color gamut boundary extraction,a color gamut boundary extraction method for digital media images based on salient region was studied in this paper.By using simple linear iterative clustering algorithm,the original digital media image was segmented by superpixel.The K-means clustering method based on texture features was used to correct the under-segmented region in the result of superpixel segmentation.The color histogram global comparison algorithm was used to calculate the color significance value of the corrected superpixel layer,and the significance region of the digital media image was extracted by combining with the Bayesian framework.The color gamut boundaries of digital media images were extracted within the salient region by means of hue angle clustering.Experimental results show that this method can effectively segment digital media images by superpixel and obtain significant regions.The color gamut boundary of digital media image can be extracted effectively.The spatial distribution of color gamut boundary points extracted by this method is more uniform,that is,the integrity of color gamut boundary extraction is better.

关 键 词：显著性区域 数字媒体图像 色域边界 提取方法 贝叶斯 色相角聚类 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]