基于图像结构的无参考模糊图像及其复原图像质量评价方法

No-Reference Blurred Image Based on Image Structure and the Quality Assessment Method of Its Restoration Image

针对传统图像的质量评价测度,提出了一种基于边缘检测的无参考模糊图像评价模型。通过测量梯度图像的梯度边缘信息,来评价模糊退化图像和带有模糊效应的复原图像,实验表明:利用本测度所得到的模糊图像质量评价结果与主观评价结果有很好的相关性。本测度较准确地反映了人对模糊图像质量的主观感受,对于带有模糊效应的复原图像的评价结果也较为符合主观评价结果。

Aiming at the traditional image quality assessment metric, this paper put forward a kind of no-reference blurred image assessment model based on edge detection. It assessed the blurred image and restoration image with blurring effect through measuring gradient edge information of gradient image. The experiment showed that the acquired quality assessment result of blurred image using this metric has good correlation with subjective assessment result. This metric truly reflected the people＇s subjective experience on the quality of blurred image. The assessment result of restoration image with blurring effect also well accorded with subjective assessment result.