【摘要】 传统的图像质量评价算法大多针对灰度图像而言,这些算法利用图像中对应的像素灰度误差建立数学模型来评价图像,不能够评价彩色图像的质量。本文提出了一种基于边缘特征和颜色亮度信息的彩色图像质量评价方法。首先,利用Sobel算子提取图像边缘;其次,定义了边缘和亮度相似系数来分析它们对图像的影响。最后,综合考虑这两个因素建立数学模型对彩色图像进行评价。大量实验表明:该方法适合评价彩色图像;同时,该实验结果与人的主观感觉的一致性较好。更多还原

【Abstract】 Traditional algorithms of image quality assessment mostly focus on gray image,they build mathematical models based on corresponding pixel gray scale error to evaluate image quality,but not to evaluate color image quality.This paper propose a new algorithm based on edge feature and color intensity information.Firstly,it extract edge feature by Sobel operatortopology;Furthermore,this algorithm define edge and intensity analogous coefficient to analyse the effects on image;Finally,this paper combine these two coefficients to assess images.Result from large amount of experiments shows that it can be used to assess color image quality;Meanwhile,the result shows good consistency between the proposed metric and subjective assessment by human observers.更多还原