基于高阶相位一致性的混合失真图像质量评价

发布时间：2018-02-20 03:24

本文关键词： 图像处理混合失真图像质量评价高阶相位一致性灰度共生矩阵相关性　出处：《激光与光电子学进展》2017年07期 　论文类型：期刊论文

【摘要】：混合失真图像质量评价是图像质量评价(IQA)领域的重点和难点,基于高阶相位一致性,提出了一种混合失真无参考IQA算法。计算了高阶相位一致性用于捕捉图像结构信息,应用灰度共生矩阵分别提取了4阶相位一致性图像的统计特征;在分析相邻阶相位一致性的相关性及相邻阶相位一致性局部熵的相关性的基础上,分别计算了相邻阶相位一致性及其局部熵的互信息和交叉熵;利用支持向量回归机制建立回归模型并进行质量预测。在MLIVE和MDID2013数据库上的实验结果表明,该算法的评价结果与主观评价分数具有很高的一致性,其性能优于当今主流的全参考和无参考IQA算法。
[Abstract]:Mixed distortion image quality evaluation is an important and difficult problem in the field of image quality evaluation. Based on high order phase consistency, a hybrid distortion without reference IQA algorithm is proposed. High order phase consistency is calculated to capture image structure information. The statistical features of the fourth order phase consistency images are extracted by using the gray level co-occurrence matrix, and the correlation of the adjacent order phase consistency and the local entropy of the adjacent order phase consistency is analyzed based on the analysis of the correlation between the adjacent order phase consistency and the adjacent order phase consistency local entropy. The mutual information and cross-entropy of adjacent order phase consistency and local entropy are calculated respectively, and the regression model and quality prediction are established by using support vector regression mechanism. The experimental results on MLIVE and MDID2013 databases show that, The evaluation results of the algorithm are consistent with the subjective evaluation scores, and its performance is better than that of the current mainstream full-reference and no-reference IQA algorithms.
【作者单位】：天津大学电气自动化与信息工程学院;
【基金】：国家自然科学基金面上项目(61471262) 重点国际(地区)合作研究项目(61520106002) 教育部博士点基金(20130032110010)
【分类号】：TP391.41

【相似文献】