摘要
主流图像质量评价算法大多基于空间域分析图像质量,当应用于彩色数字图像时,准确度和敏感度并未达到很好的效果,提出了一种基于图像频率域直方图和HVS(人类视觉系统)的彩色图像质量评价算法。将图像转换成频率空间,并对频率域系数幅度进行直方图计算;结合人类视觉对亮度和颜色的感知特点,对彩色图像的亮度和颜色分配权重并最终融合评价结果。实验结果表明,本算法与真实感官感受具有高度一致性,准确度和敏感度优于其他主流算法,同时具有最快的计算处理效率。
For the problem that the popular IQA methods do not have good performance on the measurement of color image quality assessment, a new method is proposed which based on histogram and HVS. The images transformed to frequency space and com- pute the histogram; combined with human vision system, considering the human eye＇ s different response to the luminance and col- or, compute the weighted result. Experimental result shows that this method is similar with the true human seusing result, which is better than other popular method, at the same time, it has less processing time.
出处
《电视技术》
北大核心
2015年第15期28-33,共6页
Video Engineering
基金
湖北省教育厅教育规划项目(2013B308)
长江大学工程技术学院科研发展基金项目(13J0802)
