光学与光电技术, 2014, 12 (1): 63, 网络出版: 2014-05-04  

一种无参考型模糊图像质量评价方法

A No-Reference Quality Assessment Method for Blurred Image
作者单位
1 河海大学计算机与信息学院, 江苏 常州 213022
2 常州市传感网与环境感知重点实验室, 江苏 常州 213022
摘要
图像的边缘信息是人眼观察和识别物体的重要特征,根据模糊图像相对于清晰图像其边缘特征发生较大变化的特点,提出了一种基于边缘锐度的无参考模糊图像质量评价方法。首先,通过文中所示方法寻找图像中的所有阶跃边缘;其次,根据一些原则选择合适的部分边缘;最终,计算这些合适边缘的锐利程度作为图像的模糊度评价依据。实验结果表明,该方法相比于全参考模型SSIM能够更好地评价高斯模糊、离焦模糊等模糊类型图像,与主观评价结构相关性强,更符合人眼视觉系统特性,并且易于实现。
Abstract
The edge imformation of images is an important feature for the human eyes to observe and recognise objects. According to the feature that the characteristics of the edges of blurred image will change more sharpnessly compared to the edges of clear image, a no-reference quality assessment method for blurred image based on the edge sharpness is presented in this paper. Firstly, search all step edges of the image with the approach which is presented in the paper. In addition, according to some principles the proper step edges are chosen. Finally, the sharpness of these appropriate edges are computed as the metric to access blurred images. Experimental results show that this method is able to access blurred images better than SSIM such as gauss fuzzy, defocus fuzzy and so on, and it has strong correlation with subjective evaluation method, more consistent with human visual system and easy to realize.

张伟, 糜靖峰, 李庆武, 霍冠英. 一种无参考型模糊图像质量评价方法[J]. 光学与光电技术, 2014, 12(1): 63. ZHANG Wei, MI Jing-feng, LI Qing-wu, HUO Guan-ying. A No-Reference Quality Assessment Method for Blurred Image[J]. OPTICS & OPTOELECTRONIC TECHNOLOGY, 2014, 12(1): 63.

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!