基于概率矩阵分解的多失真图像质量评估算法
|
摘要
图像质量评价(Image Quality Assessment, IQA)是视觉感知模型研究的重要分支,使用IQA算法自动化地评估图像质量有广阔的应用前景。基于概率矩阵分解(Probability Matrix Factorization)提出一种多失真图像质量评价算法,主要贡献有:提出一种图像质量评价的新思路,即利用PMF方法从失真图像估计参考图像,把失真图像和估计参考图像之间的信息损失作为图像质量的度量;构建新颖的特征向量描述这种信息损失;使用支持向量回归(Support Vector Regression, SVR)完成图像质量模型的训练。提出的算法在多个公开的图像质量评价数据库上超过了经典方法,实验结果证明该方法与人的主观质量评价具有更好的一致性。
Image quality assessment is an important branch of visual perception modeling research. Using IQA algorithm to automatically evaluate the image quality has broad application prospects. In this paper, we proposed a multiply-distorted IQA method based on probability matrix factorization. The main contributions are as follows: firstly, we came up with a new idea for IQA, which used the PMF method to estimate the reference image from the distorted image, and the information loss between the distorted image and the estimated reference version was considered as a quality measure. Secondly, we constructed a novel feature vector to describe this information loss. Finally, the IQA model was trained with SVR. The proposed method is superior to the classical IQAs in several public image quality assessment databases, which demonstrates our method has the better performance in terms of the coherence with human subjective rating.
Image quality assessment is an important branch of visual perception modeling research. Using IQA algorithm to automatically evaluate the image quality has broad application prospects. In this paper, we proposed a multiply-distorted IQA method based on probability matrix factorization. The main contributions are as follows: firstly, we came up with a new idea for IQA, which used the PMF method to estimate the reference image from the distorted image, and the information loss between the distorted image and the estimated reference version was considered as a quality measure. Secondly, we constructed a novel feature vector to describe this information loss. Finally, the IQA model was trained with SVR. The proposed method is superior to the classical IQAs in several public image quality assessment databases, which demonstrates our method has the better performance in terms of the coherence with human subjective rating.
引文
[1]周景超,戴汝为,肖柏华.图像质量评价研究综述[J].计算机科学,2008,35(7):1-4.
[2]李爱华.基于相位的结构相似度图像质量评价模型[J].计算机应用与软件,2014,31(11):233-236.
[3]Wang Z,Bovik A C,Sheikh H R,et al.Image Quality Assessment:From Error Visibility to Structural Similarity[J].IEEE Transactions on Image Processing,2004,13(4):600-612.
[4]Sheikh H R,Bovik A C.Image information and visual quality[J].IEEE Transactions on Image Processing,2006,15(2):430-444.
[5]Demirtas A M,Reibman A R,Jafarkhani H.Full-Reference Quality Estimation for Images With Different Spatial Resolutions[J].Image Processing IEEE Transactions on,2014,23(5):2069-2080.
[6]Di Claudio E D,Jacovitti G.A Detail Based Method for Linear Full Reference Image Quality Prediction[J].IEEE Trans Image Process,2018,27(1):179-193.
[7]Soundararajan R,Bovik A C.RRED Indices:Reduced Reference Entropic Differencing for Image Quality Assessment[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2012,21(2):517-526.
[8]Mittal A,Moorthy A K,Bovik A C.No-Reference Image Quality Assessment in the Spatial Domain[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2012,21(12):4695.
[9]Gu K,Zhai G,Liu M,et al.FISBLIM:A FIve-Step BLInd Metric for quality assessment of multiply distorted images[C]//2013 IEEE Workshop on Signal Processing Systems(Si PS).IEEE,2013.
[10]Chandler D M.Seven Challenges in Image Quality Assessment:Past,Present,and Future Research[J].Isrn Signal Processing,2013,2013:Article ID 905685.
[11]Ghadiyaram D,Bovik A C.Perceptual Quality Prediction on Authentically Distorted Images Using a Bag of Features Approach[J].Journal of Vision,2017,17(1):32.
[12]Li Q,Lin W,Fang Y.No-Reference Quality Assessment for Multiply-Distorted Images in Gradient Domain[J].IEEESignal Processing Letters,2016,23(4):541-545.
[13]Mittal A,Fellow,IEEE,et al.Making a’Completely Blind’Image Quality Analyzer[J].IEEE Signal Processing Letters,2013,20(3):209-212.
[14]Zhang L,Zhang L,Bovik A C.A feature-enriched completely blind image quality evaluator[J].IEEE Transactions on Image Processing,2015,24(8):2579-2591.
[15]Saha A,Wu Q M J.Utilizing image scales towards totally training free blind image quality assessment[J].IEEETransactions on Image Processing,2015,24(6):1879-1892.
[16]Zhang Y,Chandler D M.Opinion-Unaware Blind Quality Assessment of Multiply and Singly Distorted Images via Distortion Parameter Estimation[J].IEEE Transactions on Image Processing,2018,27(11):5433-5448.
[17]Mnih A,Salakhutdinov R R.Probabilistic matrix factorization[C]//Advances in neural information processing systems.2008:1257-1264.
[18]Wang N,Yeung D Y.Bayesian robust matrix factorization for image and video processing[C]//Proceedings of the IEEE International Conference on Computer Vision.2013:1785-1792.
[19]Li S,Zhang F,Ma L,et al.Image quality assessment by separately evaluating detail losses and additive impairments[J].IEEE Transactions on Multimedia,2011,13(5):935-949.
[20]Dabov K,Foi A,Katkovnik V,et al.Image denoising by sparse 3-D transform-domain collaborative filtering[J].IEEE Transactions on image processing,2007,16(8):2080-2095.
[21]Ye P,Kumar J,Kang L,et al.Unsupervised feature learning framework for no-reference image quality assessment[C]//2012 IEEE conference on computer vision and pattern recognition.IEEE,2012:1098-1105.
[22]刘国军,高丽霞,陈丽奇.广义平均的全参考型图像质量评价池化策略[J].光学精密工程,2017,25(3):742-748.
[23]Saad M A,Bovik A C,Charrier C.Blind image quality assessment:A natural scene statistics approach in the DCT domain[J].IEEE transactions on Image Processing,2012,21(8):3339-3352.
[24]付燕,张妍.基于ε-SVR的JPEG无参考图像质量评价[J].计算机应用与软件,2014,31(10):213-215,238.
[25]高飞,高新波.主动特征学习及其在盲图像质量评价中的应用[J].计算机学报,2014,37(10):2227-2234.
[26]Chang C C,Lin C J.LIBSVM:A library for support vector machines[J].ACM Transactions on Intelligent Systems and Technology,2011,2(3):1-27.
[27]Jayaraman D,Mittal A,Moorthy A K,et al.Objective quality assessment of multiply distorted images[C]//Signals,Systems and Computers(ASILOMAR),2012 Conference Record of the Forty Sixth Asilomar Conference on.IEEE,2012.
[28]Gu K,Zhai G,Yang X,et al.Hybrid No-Reference Quality Metric for Singly and Multiply Distorted Images[J].IEEETransactions on Broadcasting,2014,60(3):555-567.
[29]Xue W,Mou X,Zhang L,et al.Blind image quality assessment using joint statistics of gradient magnitude and Laplacian features[J].IEEE Transactions on Image Processing,2014,23(11):4850-4862.
[30]Saad M A,Bovik A C,Charrier C.Blind image quality assessment:A natural scene statistics approach in the DCT domain[J].IEEE transactions on Image Processing,2012,21(8):3339-3352.
[31]Li Q,Lin W,Xu J,et al.Blind image quality assessment using statistical structural and luminance features[J].IEEETransactions on Multimedia,2016,18(12):2457-2469.
[32]Gu K,Zhai G,Yang X,et al.Using free energy principle for blind image quality assessment[J].IEEE Transactions on Multimedia,2015,17(1):50-63.
[33]Hadizadeh H,Baji?I V.Color Gaussian Jet features for noreference quality assessment of multiply-distorted images[J].IEEE Signal Processing Letters,2016,23(12):1717-1721.
[2]李爱华.基于相位的结构相似度图像质量评价模型[J].计算机应用与软件,2014,31(11):233-236.
[3]Wang Z,Bovik A C,Sheikh H R,et al.Image Quality Assessment:From Error Visibility to Structural Similarity[J].IEEE Transactions on Image Processing,2004,13(4):600-612.
[4]Sheikh H R,Bovik A C.Image information and visual quality[J].IEEE Transactions on Image Processing,2006,15(2):430-444.
[5]Demirtas A M,Reibman A R,Jafarkhani H.Full-Reference Quality Estimation for Images With Different Spatial Resolutions[J].Image Processing IEEE Transactions on,2014,23(5):2069-2080.
[6]Di Claudio E D,Jacovitti G.A Detail Based Method for Linear Full Reference Image Quality Prediction[J].IEEE Trans Image Process,2018,27(1):179-193.
[7]Soundararajan R,Bovik A C.RRED Indices:Reduced Reference Entropic Differencing for Image Quality Assessment[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2012,21(2):517-526.
[8]Mittal A,Moorthy A K,Bovik A C.No-Reference Image Quality Assessment in the Spatial Domain[J].IEEE Transactions on Image Processing A Publication of the IEEE Signal Processing Society,2012,21(12):4695.
[9]Gu K,Zhai G,Liu M,et al.FISBLIM:A FIve-Step BLInd Metric for quality assessment of multiply distorted images[C]//2013 IEEE Workshop on Signal Processing Systems(Si PS).IEEE,2013.
[10]Chandler D M.Seven Challenges in Image Quality Assessment:Past,Present,and Future Research[J].Isrn Signal Processing,2013,2013:Article ID 905685.
[11]Ghadiyaram D,Bovik A C.Perceptual Quality Prediction on Authentically Distorted Images Using a Bag of Features Approach[J].Journal of Vision,2017,17(1):32.
[12]Li Q,Lin W,Fang Y.No-Reference Quality Assessment for Multiply-Distorted Images in Gradient Domain[J].IEEESignal Processing Letters,2016,23(4):541-545.
[13]Mittal A,Fellow,IEEE,et al.Making a’Completely Blind’Image Quality Analyzer[J].IEEE Signal Processing Letters,2013,20(3):209-212.
[14]Zhang L,Zhang L,Bovik A C.A feature-enriched completely blind image quality evaluator[J].IEEE Transactions on Image Processing,2015,24(8):2579-2591.
[15]Saha A,Wu Q M J.Utilizing image scales towards totally training free blind image quality assessment[J].IEEETransactions on Image Processing,2015,24(6):1879-1892.
[16]Zhang Y,Chandler D M.Opinion-Unaware Blind Quality Assessment of Multiply and Singly Distorted Images via Distortion Parameter Estimation[J].IEEE Transactions on Image Processing,2018,27(11):5433-5448.
[17]Mnih A,Salakhutdinov R R.Probabilistic matrix factorization[C]//Advances in neural information processing systems.2008:1257-1264.
[18]Wang N,Yeung D Y.Bayesian robust matrix factorization for image and video processing[C]//Proceedings of the IEEE International Conference on Computer Vision.2013:1785-1792.
[19]Li S,Zhang F,Ma L,et al.Image quality assessment by separately evaluating detail losses and additive impairments[J].IEEE Transactions on Multimedia,2011,13(5):935-949.
[20]Dabov K,Foi A,Katkovnik V,et al.Image denoising by sparse 3-D transform-domain collaborative filtering[J].IEEE Transactions on image processing,2007,16(8):2080-2095.
[21]Ye P,Kumar J,Kang L,et al.Unsupervised feature learning framework for no-reference image quality assessment[C]//2012 IEEE conference on computer vision and pattern recognition.IEEE,2012:1098-1105.
[22]刘国军,高丽霞,陈丽奇.广义平均的全参考型图像质量评价池化策略[J].光学精密工程,2017,25(3):742-748.
[23]Saad M A,Bovik A C,Charrier C.Blind image quality assessment:A natural scene statistics approach in the DCT domain[J].IEEE transactions on Image Processing,2012,21(8):3339-3352.
[24]付燕,张妍.基于ε-SVR的JPEG无参考图像质量评价[J].计算机应用与软件,2014,31(10):213-215,238.
[25]高飞,高新波.主动特征学习及其在盲图像质量评价中的应用[J].计算机学报,2014,37(10):2227-2234.
[26]Chang C C,Lin C J.LIBSVM:A library for support vector machines[J].ACM Transactions on Intelligent Systems and Technology,2011,2(3):1-27.
[27]Jayaraman D,Mittal A,Moorthy A K,et al.Objective quality assessment of multiply distorted images[C]//Signals,Systems and Computers(ASILOMAR),2012 Conference Record of the Forty Sixth Asilomar Conference on.IEEE,2012.
[28]Gu K,Zhai G,Yang X,et al.Hybrid No-Reference Quality Metric for Singly and Multiply Distorted Images[J].IEEETransactions on Broadcasting,2014,60(3):555-567.
[29]Xue W,Mou X,Zhang L,et al.Blind image quality assessment using joint statistics of gradient magnitude and Laplacian features[J].IEEE Transactions on Image Processing,2014,23(11):4850-4862.
[30]Saad M A,Bovik A C,Charrier C.Blind image quality assessment:A natural scene statistics approach in the DCT domain[J].IEEE transactions on Image Processing,2012,21(8):3339-3352.
[31]Li Q,Lin W,Xu J,et al.Blind image quality assessment using statistical structural and luminance features[J].IEEETransactions on Multimedia,2016,18(12):2457-2469.
[32]Gu K,Zhai G,Yang X,et al.Using free energy principle for blind image quality assessment[J].IEEE Transactions on Multimedia,2015,17(1):50-63.
[33]Hadizadeh H,Baji?I V.Color Gaussian Jet features for noreference quality assessment of multiply-distorted images[J].IEEE Signal Processing Letters,2016,23(12):1717-1721.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |