基于局部标准差与显著图的模糊图像质量评价方法

No-reference blurring image quality assessment based on local standard deviation and saliency map

为了有效地对模糊图像的质量进行评价,提出了基于局部标准差和显著图的无参考模糊图像质量评价方法。首先,针对待评价图像,利用高斯低通滤波器对其进行模糊化来构造参考图像。然后,利用图像的局部标准差和显著图两个特征在模糊化前后的变化情况,对原模糊图像的质量进行评价。最后分别在LIVE图像库和CSIQ图像库上对本文方法进行了验证,其中Pearson线性相关系数(PLCC)值分别达到了0.9315和0.9254,Spearman秩相关系数(SROCC)值分别达到了0.9258和0.8962。实验结果表明本文方法与当前公认性能优越的算法LPC-SI表现接近,且计算复杂度较低,耗时仅为其4.7%。

In order to evaluate the quality of blur image effectively,a no-reference image quality assessment method based on local standard deviation and saliency map is proposed.First,the Gaussian low-pass filter is used to construct a reference image through blurring the given image.Then,two features,namely local standard deviation map and saliency map,are selected to evaluate the quality of the blur image according to the changes of the two features before and after the blurring process.The proposed method is tested on LIVE database and CSIQ database,on which the Pearson linear correlation coefficients are 0.9315 and 0.9254 respectively,and the Spearman rank correlation coefficients are 0.9258 and 0.8962 respectively.Experimental results indicate that the proposed method is close to the state-of-art method LPC-SI,however its computational complexity is much lower,only about 4.7% of that of LPC-SI.