摘要
提出一种基于对比度敏感度的无参考图像清晰度评价方法,用于评价光测设备无参考电视图像的清晰度。首先通过低通滤波器构造原始图像的参考图像,并对两者进行离散余弦变换(DCT),其次将变换后的中频系数和高频系数各分为高低两档后进行离散余弦反变换处理得到各自子图像,然后计算相应子图像的结构相似度SSIM,最后将各子图像结构相似度加权求和得到图像的清晰度;为体现对比度敏感度函数(CSF)特性,权值设为该频带内所有频率系数绝对值的均值。对145幅高斯模糊图像的评价结果显示,相关系数(CC)为0.915 2,Spearman等级相关系数(SROCC)为0.907 9,离出率(OR)为0.034 5,准确性、单调性、一致性较高,可以在序列离焦电视图像中快速准确地找出离焦量最小图像,且效果优于运用较广泛的调焦评价函数梯度能量函数等4种传统梯度评价函数;对于不同类型的电视模糊图像的评价结果更加符合人眼的视觉特性,可以满足对光测设备无参考电视图像清晰度评价的要求。
A no reference image sharpness assessment method based on the property of Contrast Sensitivity Function（CSF） was proposed to realize the sharpness assessment of optical measurement equipment without reference television images.Firstly,a reference image for an original was constructed by a low-pass filter and both images were performed the Discrete Cosine Transform（DCT）,and intermediate frequency coefficients and high frequency coefficients are divided into two parts respectively to be performed the Inverse DCT（IDCT） to obtain sub-images.Then,the Structural Similarity（SSIM） of corresponding sub-images was calculated.Finally,the image sharpness is obtained through the weighted sum of sub-image structural similarities.The experiment results show that the proposed method can obtain the Pearson Correlation Coefficient（CC） in 0.915 2,Spearman Rank Order Correlation Coefficient CC（SROCC） in 0.907 9 and out Rate（OR） in 0.034 5,which shows very high accuracy,monotonicity and consistency for the Gaussian blur image.The smallest defocus image can be quickly and accurately identified in the sequence defocus images and the effect is better than those of other four traditional focus evaluation functions.For different types of television blur images,the assessment results is more accord with human visual characteristics.It can be applied to optical measurement equipment television system and give accurate and reliable sharpness assessment of no reference television image.
出处
《光学精密工程》
EI
CAS
CSCD
北大核心
2011年第10期2485-2493,共9页
Optics and Precision Engineering
基金
国家863高技术研究发展计划资助项目(No.2008AA7034320B)
关键词
图像质量评价
无参考
对比度敏感度
人眼视觉系统
清晰度
结构相似度
image quality assessment
no reference
contrast sensitivity
human visual system
sharpness
Structural Similarity（SSIM）
