基于雾密度感知增强的去雾图像质量评价方法

doi:10.19678/j.issn.1000-3428.0066460

摘要/Abstract

摘要： 去雾图像质量评价方法在图像去雾算法研究中发挥着重要作用，目前去雾质量评价方法存在雾密度感知能力不足的问题。提出一种基于雾密度感知增强的去雾图像质量评价（FDP-DIQA）方法，针对自然雾图的成像特点，从对比度信息及色彩信息2个方面提取6个特征，并结合以往的图像去雾研究结果提取3个特征，构成完整的去雾感知特征，增强对雾密度的感知能力。随后，结合雾图与去雾图像之间全局和低对比度区域的结构特征及块效应特征，对经平均池化形成的22维特征采用LightGBM进行模型构建，实现去雾图像质量评价。实验结果表明，FDP-DIQA方法在合成雾图数据集和自然雾图数据集上的表现优异，其加权平均后的斯皮尔曼等级相关系数、皮尔森线性相关系数、均方根误差分别为0.962 3、0.958 3、2.098 1，大幅领先于同类方法，与人类主观评价有较高的一致性。

关键词: 去雾图像质量评价, 图像对比度, 色彩信息, LightGBM模型, 图像去雾

Abstract: The Dehazed Image Quality Assessment（DIQA） method plays an important role in research on the image dehazing algorithm.At present，the dehazing quality assessment method has the problem of insufficient fog density perceived ability.In this regard，a new DIQA method called FDP-DIQA is proposed.According to the imagery characteristics of natural fog images，six features are extracted from the two aspects of contrast information and color information，and three features are extracted from the previous image dehazing research results to form a complete dehazing perception feature and enhance the perceived ability of fog density.Then，combined with the structural features and block effect features of the global and low-contrast regions between the fog map and defogged image，LightGBM is used for model construction of the 22-dimensional features formed by average pooling to realize quality evaluation of the defogged image. Experimental results show that the FDP-DIQA method performs well on both synthetic and natural fog map datasets，and its weighted average Spearman's Rank Correlation Coefficient（SRCC），Pearson Linear Correlation Coefficient（PLCC），and Root Mean Square Error（RMSE） are 0.962 3，0.958 3，and 2.098 1，respectively， which is a significant improvement over existing similar methods and has a high consistency with human subjective evaluation.

Key words: Dehazed Image Quality Assessment（DIQA）, image contrast, color-aware features, LightGBM model, image dehazing

中图分类号:

TP391.41

朱钦权, 王同罕, 贾惠珍. 基于雾密度感知增强的去雾图像质量评价方法[J]. 计算机工程, 2023, 49(6): 208-216,226.

ZHU Qinquan, WANG Tonghan, JIA Huizhen. Dehazed Image Quality Assessment Method Based on Fog Density Perception Enhancement[J]. Computer Engineering, 2023, 49(6): 208-216,226.

http://www.ecice06.com/CN/Y2023/V49/I6/208

图/表 12

20230615170056

20230615170059

20230615170102

20230615170105

20230615170109

20230615170112

20230615170115

20230615170118

20230615170127

20230615170131

20230615170135

20230615170138

参考文献

[1] VAN OLDENBORGH G J,YIOU P,VAUTARD R.On the roles of circulation and aerosols in the decline of mist and dense fog in Europe over the last 30 years[J].Atmospheric Chemistry and Physics,2010,10(10):4597-4609.
[2] LI B Y,PENG X L,WANG Z Y,et al.AOD-net:all-in-one dehazing network[C]//Proceedings of IEEE International Conference on Computer Vision.Washington D.C.,USA:IEEE Press,2017:4780-4788.
[3] CAI B L,XU X M,JIA K,et al.DehazeNet:an end-to-end system for single image haze removal[J].IEEE Transactions on Image Processing,2016,25(11):5187-5198.
[4] WANG Z,SIMONCELLI E P,BOVIK A C.Multiscale structural similarity for image quality assessment[C]//Proceedings of the 37th Asilomar Conference on Signals,Systems & Computers.Washington D.C.,USA:IEEE Press,2004:1398-1402.
[5] ENGIN D,GENC A,EKENEL H K.Cycle-dehaze:enhanced CycleGAN for single image dehazing[C]//Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops.Washington D.C.,USA:IEEE Press,2018:938-9388.
[6] HAUTIÈRE N,TAREL J P,AUBERT D,et al.Blind contrast enhancement assessment by gradient ratioing at visible edges[J].Image Analysis & Stereology,2011,27(2):87-92.
[7] CHOI L K,YOU J,BOVIK A C.Referenceless prediction of perceptual fog density and perceptual image defogging[J].IEEE Transactions on Image Processing,2015,24(11):3888-3901.
[8] JIANG Y T,SUN C M,ZHAO Y,et al.Fog density estimation and image defogging based on surrogate modeling for optical depth[J].IEEE Transactions on Image Processing,2017,26(7):3397-3409.
[9] GUAN T X,LI C F,GU K,et al.Visibility and distortion measurement for no-reference dehazed image quality assessment via complex contourlet transform[J].IEEE Transactions on Multimedia,2022,99(3):11-21.
[10] ZHAO S Y,ZHANG L,HUANG S Y,et al.Dehazing evaluation:real-world benchmark datasets,criteria,and baselines[J].IEEE Transactions on Image Processing,2020,29:6947-6962.
[11] ZHANG L,SHEN Y,LI H Y.VSI:a visual saliency-induced index for perceptual image quality assessment[J].IEEE Transactions on Image Processing,2014,23(10):4270-4281.
[12] ZHANG L,ZHANG L,MOU X Q,et al.FSIM:a feature similarity index for image quality assessment[J].IEEE Transactions on Image Processing,2011,20(8):2378-2386.
[13] LIU W,ZHOU F,LU T,et al.Image defogging quality assessment:real-world database and method[J].IEEE Transactions on Image Processing,2021,30(7):176-190.
[14] MIN X K,ZHAI G T,GU K,et al.Objective quality evaluation of dehazed images[J].IEEE Transactions on Intelligent Transportation Systems,2019,20(8):2879-2892.
[15] KE G L,MENG Q,FINLEY T,et al.LightGBM:a highly efficient gradient boosting decision tree[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems.New York,USA:ACM Press,2017:3149-3157.
[16] JIA H Z,ZHANG L,WANG T H.Contrast and visual saliency similarity-induced index for assessing image quality[J].IEEE Access,2018,6:65885-65893.
[17] ORTIZ-JARAMILLO B,KUMCU A,PLATISA L,et al.Content-aware contrast ratio measure for images[J].Signal Processing:Image Communication,2018,62(3):51-63.
[18] GROEN I I,GHEBREAB S,PRINS H,et al.From image statistics to scene gist:evoked neural activity reveals transition from low-level natural image structure to scene category[J].The Journal of Neuroscience:the Official Journal of the Society for Neuroscience,2013,33(48):18814-18824.
[19] SCHOLTE H S,GHEBREAB S,WALDORP L,et al.Brain responses strongly correlate with Weibull image statistics when processing natural images[J].Journal of Vision,2009,9(4):29-35.
[20] VAZQUEZ-CORRAL J,GALDRAN A,CYRIAC P,et al.A fast image dehazing method that does not introduce color artifacts[J].Journal of Real-Time Image Processing,2020,17(3):607-622.
[21] HE K M,SUN J,TANG X O.Single image haze removal using dark channel prior[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(12):2341-2353.
[22] ZHU Q S,MAI J M,SHAO L.A fast single image haze removal algorithm using color attenuation prior[J].IEEE Transactions on Image Processing,2015,24(11):3522-3533.
[23] BERMAN D,TREIBITZ T,AVIDAN S.Non-local image dehazing[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C.,USA:IEEE Press,2016:1674-1682.
[24] MIN X K,MA K D,GU K,et al.Unified blind quality assessment of compressed natural,graphic,and screen content images[J].IEEE Transactions on Image Processing,2017,26(11):5462-5474.
[25] MIN X,GU K,ZHAI G,et al.Saliency-induced reduced-reference quality index for natural scene and screen content images[J].Signal Processing,2018,145:127-136.
[26] ZOU Z Y,WANG L,CHEN J,et al.Research on peanut variety classification based on hyperspectral image[J].Food Science and Technology,2022,42(3):32-41.
[27] SHEIKH H R,SABIR M F,BOVIK A C.A statistical evaluation of recent full reference image quality assessment algorithms[J].IEEE Transactions on Image Processing 2006,15(11):3440-3451.
[28] TAREL J P,HAUTIÈRE N,CORD A,et al.Improved visibility of road scene images under heterogeneous fog[C]//Proceedings of IEEE Intelligent Vehicles Symposium.Washington D.C.,USA:IEEE Press,2010:478-485.
[29] 王效灵,胡志杰,徐帅帅,等.改进暗通道先验和策略性融合的图像去雾算法[J/OL].计算机工程,1-12[2023-01-04].https://doi.org/10.19678/j.issn.1000-3428.0065908.(in Chinese)WANG X L,HU Z J,XU S S,et al.Image dehazing algorithm by improving dark channel prior and strategic fusion[J/OC].Computer Engineering,1-12[2023-01-04].https://doi.org/10.19678/j.issn.1000-3428.0065908.
[30] 陈志华,杜垣均,温阳,等.基于注意力机制的密集残差融合与空间局部滤波低光照去雾算法[J].计算机辅助设计与图形学学报,2022,34(12):1842-1849.CHEN Z H,DU Y J,WEN Y,et al.Dense residual fusion based on attention mechanism and spatial local filtering low-light defogging algorithm[J].Journal of Computer-Aided Design & Computer Graphics,2022,34(12):1842-1849.(in Chinese)

选择文件类型/文献管理软件名称

选择包含的内容