计算机科学与探索 ›› 2017, Vol. 11 ›› Issue (4): 633-642.DOI: 10.3778/j.issn.1673-9418.1512042

• 人工智能与模式识别 • 上一篇    下一篇

四元数小波变换的无参考图像质量评价

桑庆兵+,高  双   

  1. 江南大学 物联网工程学院,江苏 无锡 214122
  • 出版日期:2017-04-12 发布日期:2017-04-12

No-Reference Image Quality Assessment via Quaternion Wavelet Transform

SANG Qingbing+, GAO Shuang   

  1. School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • Online:2017-04-12 Published:2017-04-12

摘要: 提出了一种新的利用四元数小波变换的通用型无参考图像质量评价算法,其利用四元小波金字塔将二维图像映射到四维空间,每层可以表示为1个幅值和3个相位(Φθψ),其中ψ相位包含丰富的图像纹理信息,能有效表征图像的结构信息。因此,通过提取各尺度ψ相位中能有效表征图像失真程度的特征,并构成特征向量,通过支持向量回归(support vector regression,SVR)模型预测图像质量得分。实验结果表明,该算法能有效反映各失真类型图像的视觉感知质量,斯皮尔曼等级相关系数值能达到0.942。

关键词: 无参考图像质量评价, 四元数小波变换, 四元小波金字塔, 支持向量回归(SVR)

Abstract: This paper proposes a new general purpose no-reference image quality assessment based on quaternion wavelet transform. It projects a two-dimenisonal image to four-dimenisonal space by using the quaternion wavelet pyramid. Each level of pyramid provides a shift-invariant magnitude and 3-angle (Φ, θ, ψ) phase. Phase ψ includes image texture information and can effectively characterize the structural information of the image. So the method extracts the features which can reflect the degree of image distortion and constitute the feature vector. Finally, the extracted feature is inputted to the support vector regression (SVR) model to predict image quality score. The experimental results show that the proposed algorithm can effectively reflect the visual quality of the image in different distortion types. Spearman correlation coefficient (SROCC) value can reach 0.942.

Key words:  no-reference image quality assessment, quaternion wavelet transform, quaternion wavelet pyramid, support vector regression (SVR)