CAJ | 학술논문

提出了一种基于小波变换和人类视觉系统的图像质量评价方法。首先分别测得4级小波分解后高低频分量的 SSIM 值并用 CSF 曲线加权，然后对低频分量进行分块 DCT 变换，测得每块的中高频分量的 SSIM 值作为乘性系数与前面所得结果相乘，最后对彩色图像各个通道分别进行加权相加。实验结果表明，算法与人眼主观感受值更加吻合。
제출료일충기우소파변환화인류시각계통적도상질량평개방법。수선분별측득4급소파분해후고저빈분량적 SSIM 치병용 CSF 곡선가권，연후대저빈분량진행분괴 DCT 변환，측득매괴적중고빈분량적 SSIM 치작위승성계수여전면소득결과상승，최후대채색도상각개통도분별진행가권상가。실험결과표명，산법여인안주관감수치경가문합。
This paper presents an image quality assessment methods based on wavelet transform and human visual system. First of all, we measured the SSIM value of high frequency and low frequency components of four levels wavelet decomposition and weighted by CSF curve. Then we measured the SSIM value of the high frequency component and middle frequency component of each DCT blocks as a multiplicative factor. Finally, for each channel of color image respectively weighted sum. Experimental results show that this algorithm is more consistent with human subjective value.