复旦大学信息科学与工程学院智慧网络与系统研究中心和电子工程系,上海,200433
网络出版:2019-04-25,
纸质出版:2019
移动端阅览
陈扬, 李旦, 张建秋. 互补色小波域图像质量盲评价方法[J]. 电子学报, 2019,47(4):775-783.
CHEN Yang, LI Dan, ZHANG Jian-qiu. Blind Image Quality Assessment with Complementary Color Wavelet Transform[J]. Acta Electronica Sinica, 2019, 47(4): 775-783.
陈扬, 李旦, 张建秋. 互补色小波域图像质量盲评价方法[J]. 电子学报, 2019,47(4):775-783. DOI: 10.3969/j.issn.0372-2112.2019.04.002.
CHEN Yang, LI Dan, ZHANG Jian-qiu. Blind Image Quality Assessment with Complementary Color Wavelet Transform[J]. Acta Electronica Sinica, 2019, 47(4): 775-783. DOI: 10.3969/j.issn.0372-2112.2019.04.002.
图像色彩空间的RGB通道具有密切的关系,图像质量的改变会改变这样的关系.然而传统图像质量评价方法大多基于灰度图像统计特性,忽略了颜色通道间关系信息.为充分利用颜色信息,本文基于新近提出的互补色小波变换提出一种图像质量盲评价方法.文章建立了图像互补色域自然场景统计、多尺度和方向性能量分布等模型.分析表明:这些模型不仅涵盖了传统灰度方法所能描述的信息,而且还能借助于互补色来有效表示彩色图像各通道之间的信息联系,提供表征图像质量的一组高效特征.基于这些特征,我们提出的图像质量盲评价的方法能有效提取图像的失真统计特征,能给出与人眼主观评价图像质量结果保持高度一致、优于现有文献报道盲方法、且可与非盲(全参考)方法相比拟的评价结果.
In the image color space
the RGB channels have strong correlations.The quality change of an image will lead to the change of channel correlations.However
most traditional image quality assessment (IQA) methods
based on grayscale image statistics
ignore such correlation information among color channels.In this paper
to utilize the color information
we propose a blind IQA method based on the recent proposed complementary color wavelet transform (CCWT).We provide models for the complementary color nature scene statistics
multi-resolution and multi-directionality energy distributions of an image.The analysis shows that our models not only cover the information of traditional methods
but also provide the relation information among color channels.A group of high-efficiency image quality features is then given.Based on these features
our blind IQA method can effectively extract the distortion statistic features and provide assessment results.Our IQA results are agreeing with the human subjective
better than the state-of-the-art blind IQA results
and close to the full-reference ones.
0
浏览量
549
下载量
3
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621