基于掩盖效应的图像质量综合评价方法
|
摘要
在对图像质量评价方法的研究中,HVS模型充当了一个重要的角色,它的引入使得图像质量的评价结果更加符合人眼的主观感受,大大克服了传统图像质量评价方法的缺点。由于HVS模型涉及到了人眼的生理特性,给研究工作带来了一定的难度,为此,需要将HVS模型的各个组成部分分开研究。掩盖效应作为其中的一部分,它的研究是对HVS模型深入了解的必经之路,也是图像质量评价方法发展道路上的奠基石。
本文着重研究了掩盖模型,根据标准图像不同类型给出了新的图像质量评价方法,并且进一步考察了误差图像对图像质量评价模型的影响。主要研究工作及成果如下：
1.对已有的关于掩盖效应的知识归纳总结,并对每种掩盖效应的物理概念、实验参数和计算公式等内容做出了说明。
2.根据韦勃—费赫涅耳定律,就掩盖阂值的测量给出了详细的方法、步骤和结果,并将掩盖阈值的计算结果应用到掩盖效应的模型计算中。
3.针对掩盖效应的不同类型,给出了相应的图像质量评价方法,其中包含基本思路、数学模型和编程步骤,并对实验结果进行分析。
4.将空间相似性与掩盖效应相结合,提出了一种基于掩盖模型空间相似性的图像质量评价方法。该方法考察了误差图像的无关分量和相关分量,并且对两个误差分量无关分量和三个误差分量无关分量做出比较,通过实验平台的建立和对大量图像的考察,得出实验数据及分析结果。
经过大量的实验证明,本方法较之传统方法有明显提高,更加贴近人眼主观感受。
In the study of image quality evaluation method, the HVS model serves as an important position, it makes the image quality evaluation results are more consistent with human subjective feelings, greatly overcome the disadvantages of traditional image quality evaluation method. Because HVS model relates to the human physiological characteristics, it brings a certain degree of difficulty to research work. For this, we need to research each component of the HVS model separately. Masking effect as part of the HVS model, it's research is the only way which must be passed to the HVS model in-depth understanding, is also the foundation stone of image quality evaluation method development road.
This paper focuses on the study of masking model, gives new image quality assessment methods based on different types of standard image, and further investigated the impact of the error image on the image quality evaluation model. Major research work and achievements are as follows:
1. Summarize the existing knowledge on the masking effect, and make notes of each masking effect's physical concepts, experimental parameters and calculate formula and other content to.
2. According to Weber-Fechner's law, it gives a detailed methods, procedures and results of masking threshold measurements. The masking threshold calculation results are applied to the calculation of masking effect model.
3. For different types of masking effect, we have given the corresponding image quality evaluation method, which contains the basic idea, the mathematical model and the programming steps. Analyze the experimental results.
4. Combining spatial similarity and masking effect, this paper presents the image quality evaluation method based on masking model and spatial similarity. The method examines the error image's unrelated component and related component, and makes a comparison on the two error component's unrelated component and three error component's unrelated component. Through the experimental platform and a large number of investigation of image, we can get the experimental data and analysis results.
After a lot of experiments proved that this method has obviously improved, and more closer to the human subjective experience than the traditional methods.
本文着重研究了掩盖模型,根据标准图像不同类型给出了新的图像质量评价方法,并且进一步考察了误差图像对图像质量评价模型的影响。主要研究工作及成果如下：
1.对已有的关于掩盖效应的知识归纳总结,并对每种掩盖效应的物理概念、实验参数和计算公式等内容做出了说明。
2.根据韦勃—费赫涅耳定律,就掩盖阂值的测量给出了详细的方法、步骤和结果,并将掩盖阈值的计算结果应用到掩盖效应的模型计算中。
3.针对掩盖效应的不同类型,给出了相应的图像质量评价方法,其中包含基本思路、数学模型和编程步骤,并对实验结果进行分析。
4.将空间相似性与掩盖效应相结合,提出了一种基于掩盖模型空间相似性的图像质量评价方法。该方法考察了误差图像的无关分量和相关分量,并且对两个误差分量无关分量和三个误差分量无关分量做出比较,通过实验平台的建立和对大量图像的考察,得出实验数据及分析结果。
经过大量的实验证明,本方法较之传统方法有明显提高,更加贴近人眼主观感受。
In the study of image quality evaluation method, the HVS model serves as an important position, it makes the image quality evaluation results are more consistent with human subjective feelings, greatly overcome the disadvantages of traditional image quality evaluation method. Because HVS model relates to the human physiological characteristics, it brings a certain degree of difficulty to research work. For this, we need to research each component of the HVS model separately. Masking effect as part of the HVS model, it's research is the only way which must be passed to the HVS model in-depth understanding, is also the foundation stone of image quality evaluation method development road.
This paper focuses on the study of masking model, gives new image quality assessment methods based on different types of standard image, and further investigated the impact of the error image on the image quality evaluation model. Major research work and achievements are as follows:
1. Summarize the existing knowledge on the masking effect, and make notes of each masking effect's physical concepts, experimental parameters and calculate formula and other content to.
2. According to Weber-Fechner's law, it gives a detailed methods, procedures and results of masking threshold measurements. The masking threshold calculation results are applied to the calculation of masking effect model.
3. For different types of masking effect, we have given the corresponding image quality evaluation method, which contains the basic idea, the mathematical model and the programming steps. Analyze the experimental results.
4. Combining spatial similarity and masking effect, this paper presents the image quality evaluation method based on masking model and spatial similarity. The method examines the error image's unrelated component and related component, and makes a comparison on the two error component's unrelated component and three error component's unrelated component. Through the experimental platform and a large number of investigation of image, we can get the experimental data and analysis results.
After a lot of experiments proved that this method has obviously improved, and more closer to the human subjective experience than the traditional methods.
引文
[1]高新波,路文.视觉信息质量评价方法[M].西安电子科技大学出版社,2010.
[2]魏政刚,袁杰辉,蔡元龙.图像质量评价方法的历史、现状和未来[J].中国图像图形学报,1998,3(5)：386-389.
[3]李永强,沈庆国,朱江,汪莉.数字视频质量评价方法综述[J].节目制作与广播,2006,6.74-77.
[4]汪孔桥.数字图像的质量评价[J].测控技术,2000,19(5)：14-16.
[5]杨军,刘藻珍,王寅龙.基于人眼视觉特性的图像质量评价模型[J].弹箭与制导学报,2003,23(2)：59-62.
[6]魏崇奎.基于HVS图像质量评价模型的方法研究[D].长沙：国防科学技术大学,2003.
[7]丁绪星,朱日宏,李建欣.一种基于人眼视觉特性的图像质量评价[J].中国图像图形学报,2004,9(2)：190-194.
[8]盛承光.基于HVS的图像隐写术的应用与研究[D].武汉：武汉理工大学,2007.
[9]白永强,朱仲杰,段智勇,粱丰.结合HVS特性的视频图像质量评价[J].计算机工程与应用,2009,(27).
[10]肖娜.基于小波变换和HVS的图像水印算法研究[D].长沙：湖南大学,2006.
[11]Daly S. The Visible Differences predictor:An Algorithm for the Assessment of Image Fidelity, in:Watson A. B. (Ed), Digital Images and Human Vision, MIT Press, Cambridge, MA,1993, pp:179-206.
[12]Foley J.M. Human Luminance Pattern-vision mechanisms:Masking Experiments Require a New Model. Journal of optical society of America A. June 1994. vol.11, no.6, pp.1710-1719.
[13]Teo P. T, Heeger DJ. Perceptual Image Distortion, Proc. Of the SPIE,1994, vol.2179, pp.127-141.
[14]Andrew B. Watson, Robert Borthwick, Mathias Taylor. Image quality and entropy masking [J]. SPIE Proceedings,1997,3016:358-371.
[15]A Toet and M.P. Lucassen. ANew Universal Color Image Fidelity Metric [J]. Displays,2003,24(4-5):197-207.
[16]Patrick Le Callet, Dominique Barba. A robust quality metric for color image quality assessment [J]. Proceedings of the IEEE International Conference on Image Processing.2003:437-440.
[17]C.J. van den Branden Lambrecht. Color Moving Pictures Quality Metric [J]. Proceedings of the IEEE International Conference on Image Processing, 1996:885-888.
[18]王正友,肖文.基于掩盖效应的无参考数字图像评价[J].计算机应用,2006,26(12)：2838-2840.
[19]曹岭,杨小帆,柏森.一种基于视觉掩盖的数字图像水印技术[J].重庆大学学报(自然科学版),2005,28(6)：64-67.
[20]肖亮,韦志辉,吴慧中.一种利用人眼视觉掩盖的小波域数字水印[J].通信学报,2002,23(3)：100-106.
[21]刘琼,胡瑞敏,韩镇.一种利用视觉掩蔽效应的感兴趣区域视频编码方法[J].中国图象图形学报,2010,15(6)：861-866.
[22]张煜东,吴乐南,韦耿.一个新的空间掩盖效应拟合公式[J].中国光学与应用光学,2008,1(1)：70-74.
[23]孔月萍,曾平,姜楠.基于HVS的彩色逆半调色差预测及可视化方法[J].华中科技大学学报(自然科学版),2008,(08).
[24]曹圣群,黄普明,鞠德航.HVS模型及其在静止图象压缩质量评价中的应用[J].中国图象图形学报,2003,8(4)：379-385.
[25]王楠楠,李桂苓.符合人眼视觉特性的视频质量评价模型[J].中国图像图形学报,2001,6(6)：523-527.
[26]李琼.基于Internet的人眼视觉应用及特性测量的研究[D].昆明：云南师范大学,2009.
[27]黄轩.符合人眼视觉特性的视频质量评价模型[D].合肥：中国科学技术 大学,2009.
[28]张漫.图像质量评价算法的研究[D].济南：山东大学,2007.
[29]D. H. KELLY, "Visual Responses to Time-Dependent Stimuli. I. Amplitude Sensitivity Measurements," J. Opt. Soc. Am.51,422-429 (1961)
[30]俞斯乐,侯正信,冯启明,李文元.电视原理[M].北京：国防工业出版社,2000.
[31]明军.电视测量原理[M].北京：航空工业出版社,1995：65-70.
[32]肖文.空域无参考的图像质量评价[D].南昌：江西财经大学,2006.
[33]杨红梅.数字图像水印不可感知性评价方法研究[D].青岛：山东科技大学,2009.
[34]黄伟.基于人眼视觉系统的图像数字水印研究[D].南京：南京信息工程大学,2008.
[35]刘明娜.基于视觉系统和特征提取的图像质量客观评价方法及应用研究[D].上海：上海交通大学,2010.
[36]柳慧超.基于图像空间多通道相似分解的图像质量综合评价方法[D].合肥：安徽大学,2011.
[37]魏崇奎,成礼智.一种基于掩盖效应的感知域图像质量评价方法[J].中国图象图形学报,2004,9(2)：195-200.
[38]胡国胜.结合感知与结构失真信息的视频质量评价模型[D].南昌：江西财经大学,2009.
[39]王正友,黄隆华.基于对比度灵敏度的图像质量评价方法的研究[J].计算机应用,2006,26(8)：1857-1859.
[40]Ngan K. N., Rao K. S., Singh H., Cosine Transform Coding Incorpora ting Human Visual System Model. SPIE fiber'86, Cambridge,MA,U.S. A.,pp.165-171,1986.
[41]Mannos J. L., Sakrison D. J., The Effects of a Visual Fidelity Criterio n on the Encoding of Images. IEEE Trans. Information Theory, vol.20, No.4(525-536),1974.
[42]A. B. Watson. DCT quantization matrices visually optimized for individual images [A]. Proc. of SPIE Conf Human Vision, Visual Processing and Digital display (IV) [C]. London,1993,1913:202-216.
[43]A. B. Watson, G. Y. Yang, J. A. Solomon, J. Villasenor. Visibility of wavelet quantization noise [J]. IEEE Trans. Image Processing,1997,6(8):1164-1175.
[44]A. P. Bradley. A wavelet visible difference predictor [J]. IEEE Trans. Image Processing,1999,8(5):717-730.
[45]Y. K. Lai, C.-C. J. Kuo. A Haar wavelet approach to compressed image quality measurement [J]. Journal of Visual Communication and Image Representation,2000,11(1):17-40.
[2]魏政刚,袁杰辉,蔡元龙.图像质量评价方法的历史、现状和未来[J].中国图像图形学报,1998,3(5)：386-389.
[3]李永强,沈庆国,朱江,汪莉.数字视频质量评价方法综述[J].节目制作与广播,2006,6.74-77.
[4]汪孔桥.数字图像的质量评价[J].测控技术,2000,19(5)：14-16.
[5]杨军,刘藻珍,王寅龙.基于人眼视觉特性的图像质量评价模型[J].弹箭与制导学报,2003,23(2)：59-62.
[6]魏崇奎.基于HVS图像质量评价模型的方法研究[D].长沙：国防科学技术大学,2003.
[7]丁绪星,朱日宏,李建欣.一种基于人眼视觉特性的图像质量评价[J].中国图像图形学报,2004,9(2)：190-194.
[8]盛承光.基于HVS的图像隐写术的应用与研究[D].武汉：武汉理工大学,2007.
[9]白永强,朱仲杰,段智勇,粱丰.结合HVS特性的视频图像质量评价[J].计算机工程与应用,2009,(27).
[10]肖娜.基于小波变换和HVS的图像水印算法研究[D].长沙：湖南大学,2006.
[11]Daly S. The Visible Differences predictor:An Algorithm for the Assessment of Image Fidelity, in:Watson A. B. (Ed), Digital Images and Human Vision, MIT Press, Cambridge, MA,1993, pp:179-206.
[12]Foley J.M. Human Luminance Pattern-vision mechanisms:Masking Experiments Require a New Model. Journal of optical society of America A. June 1994. vol.11, no.6, pp.1710-1719.
[13]Teo P. T, Heeger DJ. Perceptual Image Distortion, Proc. Of the SPIE,1994, vol.2179, pp.127-141.
[14]Andrew B. Watson, Robert Borthwick, Mathias Taylor. Image quality and entropy masking [J]. SPIE Proceedings,1997,3016:358-371.
[15]A Toet and M.P. Lucassen. ANew Universal Color Image Fidelity Metric [J]. Displays,2003,24(4-5):197-207.
[16]Patrick Le Callet, Dominique Barba. A robust quality metric for color image quality assessment [J]. Proceedings of the IEEE International Conference on Image Processing.2003:437-440.
[17]C.J. van den Branden Lambrecht. Color Moving Pictures Quality Metric [J]. Proceedings of the IEEE International Conference on Image Processing, 1996:885-888.
[18]王正友,肖文.基于掩盖效应的无参考数字图像评价[J].计算机应用,2006,26(12)：2838-2840.
[19]曹岭,杨小帆,柏森.一种基于视觉掩盖的数字图像水印技术[J].重庆大学学报(自然科学版),2005,28(6)：64-67.
[20]肖亮,韦志辉,吴慧中.一种利用人眼视觉掩盖的小波域数字水印[J].通信学报,2002,23(3)：100-106.
[21]刘琼,胡瑞敏,韩镇.一种利用视觉掩蔽效应的感兴趣区域视频编码方法[J].中国图象图形学报,2010,15(6)：861-866.
[22]张煜东,吴乐南,韦耿.一个新的空间掩盖效应拟合公式[J].中国光学与应用光学,2008,1(1)：70-74.
[23]孔月萍,曾平,姜楠.基于HVS的彩色逆半调色差预测及可视化方法[J].华中科技大学学报(自然科学版),2008,(08).
[24]曹圣群,黄普明,鞠德航.HVS模型及其在静止图象压缩质量评价中的应用[J].中国图象图形学报,2003,8(4)：379-385.
[25]王楠楠,李桂苓.符合人眼视觉特性的视频质量评价模型[J].中国图像图形学报,2001,6(6)：523-527.
[26]李琼.基于Internet的人眼视觉应用及特性测量的研究[D].昆明：云南师范大学,2009.
[27]黄轩.符合人眼视觉特性的视频质量评价模型[D].合肥：中国科学技术 大学,2009.
[28]张漫.图像质量评价算法的研究[D].济南：山东大学,2007.
[29]D. H. KELLY, "Visual Responses to Time-Dependent Stimuli. I. Amplitude Sensitivity Measurements," J. Opt. Soc. Am.51,422-429 (1961)
[30]俞斯乐,侯正信,冯启明,李文元.电视原理[M].北京：国防工业出版社,2000.
[31]明军.电视测量原理[M].北京：航空工业出版社,1995：65-70.
[32]肖文.空域无参考的图像质量评价[D].南昌：江西财经大学,2006.
[33]杨红梅.数字图像水印不可感知性评价方法研究[D].青岛：山东科技大学,2009.
[34]黄伟.基于人眼视觉系统的图像数字水印研究[D].南京：南京信息工程大学,2008.
[35]刘明娜.基于视觉系统和特征提取的图像质量客观评价方法及应用研究[D].上海：上海交通大学,2010.
[36]柳慧超.基于图像空间多通道相似分解的图像质量综合评价方法[D].合肥：安徽大学,2011.
[37]魏崇奎,成礼智.一种基于掩盖效应的感知域图像质量评价方法[J].中国图象图形学报,2004,9(2)：195-200.
[38]胡国胜.结合感知与结构失真信息的视频质量评价模型[D].南昌：江西财经大学,2009.
[39]王正友,黄隆华.基于对比度灵敏度的图像质量评价方法的研究[J].计算机应用,2006,26(8)：1857-1859.
[40]Ngan K. N., Rao K. S., Singh H., Cosine Transform Coding Incorpora ting Human Visual System Model. SPIE fiber'86, Cambridge,MA,U.S. A.,pp.165-171,1986.
[41]Mannos J. L., Sakrison D. J., The Effects of a Visual Fidelity Criterio n on the Encoding of Images. IEEE Trans. Information Theory, vol.20, No.4(525-536),1974.
[42]A. B. Watson. DCT quantization matrices visually optimized for individual images [A]. Proc. of SPIE Conf Human Vision, Visual Processing and Digital display (IV) [C]. London,1993,1913:202-216.
[43]A. B. Watson, G. Y. Yang, J. A. Solomon, J. Villasenor. Visibility of wavelet quantization noise [J]. IEEE Trans. Image Processing,1997,6(8):1164-1175.
[44]A. P. Bradley. A wavelet visible difference predictor [J]. IEEE Trans. Image Processing,1999,8(5):717-730.
[45]Y. K. Lai, C.-C. J. Kuo. A Haar wavelet approach to compressed image quality measurement [J]. Journal of Visual Communication and Image Representation,2000,11(1):17-40.