基于非下采样Contourlet变换的无参考图像质量评价
Learning a Blind Measurement of Image Quality by Feature from Nonsubsampled Contourlet Transform
为了度量不同失真类型的图像质量，提出了一种基于非下采样Contourlet变换和广义回归神经网络的无参考图像质量评价方法。该方法首先对原始图像进行三层非下采样Contourlet分解，然后计算各个子带的能量，最后利用广义回归神经网络建立图像子带能量与图像质量之间的关系模型。实验结果表明，本文提出的方法和主观感知结果具有较好的一致性。
o estimate a range of image distortions，a novel no-reference image quality assessment method was proposed based on the nonsubsampled Contourlet transform and general regression neural network (GRNN). Firstly, we apply three-layer nonsubsampled contourlet decomposition to the original image, and then calculate the energy of each sub-band, and lastly establish the relationship between the energy of each sub-band and image quality by GRNN. Experimental results show that our method highly correlates with human perception and can be applied to different distortions.
李任、李朝锋
计算技术、计算机技术电子技术应用
无参考图像质量评价非下采样Contourle变换（NSCT）能量广义回归神经网络（GRNN）
no-reference image quality assessmentnonsubsampled contourlet transform（NSCT）energygeneral regression neural network (GRNN)
李任,李朝锋.基于非下采样Contourlet变换的无参考图像质量评价[EB/OL].(2013-03-27)[2025-08-02].http://www.paper.edu.cn/releasepaper/content/201303-906.点此复制
评论