摘要
针对目前无参考图像质量评价算法依赖训练样本集、算法复杂性较高,难以应用于图像视频采集设备等问题,提出了一种基于有效再模糊的无参考图像质量评价算法(NRIQAVR)。首先研究了便携式设备采集图像的特点,然后在再模糊理论基础上通过计算最小和最大再模糊标准差,构造了与人类视觉感知相近的NRIQAVR算法,通过理论证明该算法可以反映自然场景图像模糊度。最后在多个公开数据集和头盔摄像机采集的图像上测试,对比其他流行算法,该算法性能稳定且无需训练,评价结果与主观感知之间的相关系数平均值高于0.92,离出率低于0.05,平均执行时间0.088 s。该算法符合人类视觉感知、不依赖图像内容、简单易行、可以达到11帧/s的处理速度,具有重要实际应用价值。
Aiming at the problem that current no-reference image quality assessment algorithms rely on training sample set,have high computing complexity and are difficult to be applied in portable image acquisition equipment. A No-reference image quality assessment algorithm based on valid re-blur（ NRIQAVR） is proposed in this paper. Firstly,the image acquisition characteristic of portable equipment is studied. Then,on the basis of re-blur theory,through calculating the minimum and maximum re-blur standard deviations of valid re-blur,the NRIQAVR algorithm is constructed,which has the feature of similar to human visual perception. It is proved theoretically that the algorithm can reflect the blur extent in natural scene images and has great similarity with visual perception. Finally,experiments on both public data-set and real data-set of the images acquired with portable camera were conducted,and the proposed algorithm was tested. The results show that compared with other popular algorithms,the NRIQAVR algorithm has a steady performance and requires no training. The average correlation coefficient between the evaluation result of the NRIQAVR algorithm and subjective perception is higher than 0. 92. The outlier ratio is less than 0. 05 and the average execution time is 0. 088 s. The proposed algorithm accords with human visual perception,does not rely on the image content,is simple to be implemented and can achieve the processing speed of 11 frames per second. The proposed algorithm can be implemented on portable equipment and has important application value.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2016年第7期1647-1655,共9页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(61372046)
陕西省自然科学基金(2014JM8338)项目资助
关键词
便携式设备
图像质量评价
视觉感知
再模糊理论
portable equipment
image quality assessment
visual perception
re-blur theory
