基于颜色线索的图像彩色化研究
|
摘要
人类通过视觉观察事物,而图像和颜色是人类通过视觉感知世界万物的两个基本要素,对于人类视觉来说,彩色图像能活泼、生动、清晰的表达事物对象,而当今世界灰度图像仍被广泛应用在人类生活的各个领域,但是由于人眼对灰度层次的识别限制和灰度图像自身的信息匮乏,相比灰度图像,彩色图像包含更丰富的信息,具有更强的对比度,给人的视觉效果更加直观和容易辨识。通过给灰度图像上色达到彩色图像的效果在各领域具有重要的研究意义,尤其是在娱乐、医学、遥感、动漫等领域,本文将先进的计算机技术应用到灰度图像的彩色化技术中,对基于线索的彩色化技术和线索位置提示方法进行了研究,具体工作如下：
本文首先对经典的彩色化技术和涉及本文算法的数字图像知识进行了介绍,然后结合测地距离和不平度,对Yatziv的颜色混合彩色化算法进行改进,从线索点开始,利用灰度图像像素点的亮度信息对测地距离和不平度进行计算,通过颜色扩展将色度在整个图像中传播,最后结合色度混合的思想,将几种具有最大贡献的色度加权最终计算出像素点的色度。随后本文将改进的彩色化方法扩展,应用到彩色图像重着色问题上,基于亮度通道在彩色图像中只能提供一部分图像信息,因此本文综合利用图像的三通道,利用颜色距离来计算像素点的测地距离,将用户提供的新的颜色在图像空间中传播开,并同样的利用颜色混合的思想最终实现彩色图像重着色。试验结果表明,本文算法具有较高的效率,彩色化效果良好。
对线索位置的提示方法进行了研究,本文提出了一种针对本文彩色化算法的基于区域生长的图像区域分割方法,区域生长的准则是由种子点开始向邻域扩展,邻域像素离种子点的亮度测地距离越小,则优先扩展。对线索位置的提示方法的研究能有效提高彩色化处理的效率,保证涂色的充分性和合理性,进而减少不合理的彩色化结果。
Human beings observe things by vision and image and color are the two basic elements for them to perceive all things in the world. For human vision, the colorful image can express the objects lively, vividly and clearly. In the modern society, gray-scale images are still widely used in various fields of human life, but due to the recognition limitation of gray level by human eyes and the information lack of gray-level images themselves, colorful images include more extensive information and own more strong contrast compared to the gray-level images, which is more direct and easy to identify. It has important research significance in various areas to color the gray-level images and realize the colorful effect, especially in entertainment, medicine, remote sensing, animation and other fields. This paper applies the advanced computer technology into the color technology of gray level images and does a research on cue-based colorization technology and cue location method. The details are as follows.
The paper first introduces the classical color techniques and digital image knowledge involved in algorithms, and then improves color blending algorithm that Yatziv proposed combining geodesic distance and unevenness. Beginning from the cue point, by the use of the brightness information of the gray-level image pixels, based on the calculation of geodesic distance and unevenness, it spreads the chromaticity throughout the whole image by color expansion, and then calculates the chromaticity of the final color pixel by combining the thought of chromaticity blending and weighting several chromaticity with maximum contribution. Then the paper will extend the color approach improved in the paper to the re-coloring issue of color images. As the luminance channel can only provide part of image information in color images, by the comprehensive utilization of the three channels of images, the paper use color distance to calculate geodesic distance of pixel, extend the new color provided by the users in the image space, and finally realize re-coloring of color images by color blending. The testing results show that the algorithm has high efficiency and good color effects.
By researching the method of clue location, the paper proposes an image segmentation algorithm based on region growing for the color algorithm. The criterion of region growing is to expand from seed point to the neighborhood. The smaller geodesic distance based on brightness between neighborhood pixel and seed point is extended with priority. The research on clue method of clue location can effectively improve the efficiency of color treatment and ensure the adequate and rational coloring, thereby reducing the unreasonable color results.
本文首先对经典的彩色化技术和涉及本文算法的数字图像知识进行了介绍,然后结合测地距离和不平度,对Yatziv的颜色混合彩色化算法进行改进,从线索点开始,利用灰度图像像素点的亮度信息对测地距离和不平度进行计算,通过颜色扩展将色度在整个图像中传播,最后结合色度混合的思想,将几种具有最大贡献的色度加权最终计算出像素点的色度。随后本文将改进的彩色化方法扩展,应用到彩色图像重着色问题上,基于亮度通道在彩色图像中只能提供一部分图像信息,因此本文综合利用图像的三通道,利用颜色距离来计算像素点的测地距离,将用户提供的新的颜色在图像空间中传播开,并同样的利用颜色混合的思想最终实现彩色图像重着色。试验结果表明,本文算法具有较高的效率,彩色化效果良好。
对线索位置的提示方法进行了研究,本文提出了一种针对本文彩色化算法的基于区域生长的图像区域分割方法,区域生长的准则是由种子点开始向邻域扩展,邻域像素离种子点的亮度测地距离越小,则优先扩展。对线索位置的提示方法的研究能有效提高彩色化处理的效率,保证涂色的充分性和合理性,进而减少不合理的彩色化结果。
Human beings observe things by vision and image and color are the two basic elements for them to perceive all things in the world. For human vision, the colorful image can express the objects lively, vividly and clearly. In the modern society, gray-scale images are still widely used in various fields of human life, but due to the recognition limitation of gray level by human eyes and the information lack of gray-level images themselves, colorful images include more extensive information and own more strong contrast compared to the gray-level images, which is more direct and easy to identify. It has important research significance in various areas to color the gray-level images and realize the colorful effect, especially in entertainment, medicine, remote sensing, animation and other fields. This paper applies the advanced computer technology into the color technology of gray level images and does a research on cue-based colorization technology and cue location method. The details are as follows.
The paper first introduces the classical color techniques and digital image knowledge involved in algorithms, and then improves color blending algorithm that Yatziv proposed combining geodesic distance and unevenness. Beginning from the cue point, by the use of the brightness information of the gray-level image pixels, based on the calculation of geodesic distance and unevenness, it spreads the chromaticity throughout the whole image by color expansion, and then calculates the chromaticity of the final color pixel by combining the thought of chromaticity blending and weighting several chromaticity with maximum contribution. Then the paper will extend the color approach improved in the paper to the re-coloring issue of color images. As the luminance channel can only provide part of image information in color images, by the comprehensive utilization of the three channels of images, the paper use color distance to calculate geodesic distance of pixel, extend the new color provided by the users in the image space, and finally realize re-coloring of color images by color blending. The testing results show that the algorithm has high efficiency and good color effects.
By researching the method of clue location, the paper proposes an image segmentation algorithm based on region growing for the color algorithm. The criterion of region growing is to expand from seed point to the neighborhood. The smaller geodesic distance based on brightness between neighborhood pixel and seed point is extended with priority. The research on clue method of clue location can effectively improve the efficiency of color treatment and ensure the adequate and rational coloring, thereby reducing the unreasonable color results.
引文
[1]Ming Yang. Still Image Colorization[EB/OL]. Available:http://www.ece. northwestern.edu/-mya671/Colorization.htm,2005-3-7
[2]Marshall P. Any Color You Like [EB/OL]. Available:http://photography .about.com/library/weekly/aa061002a.htm,2002-07-11
[3]Museum of Broadcast Communication:Encyclopedia of Television [EB/OL]. Available:http://www.museum.tv/archives/etv/C/htmlC/colorization/colorization .htm,2008-06-11
[4]Sykora D, Burianek J, and ZaraJ. Unsupervised colorization of black and white cartoons[C], In:Proceedings of the 3th international symposium on Non-photo-realistic animation and rendering, Annecy, France,2004,7:21-127
[5]Daniel LR, Thomas WC, Chuan CC. Statistics of Cone Responses to Natural Images:Implications for Visual Coding[J]. Journal of the OpticM Society of America,1998,15(8):2036-2045
[6]Reinhard E, AshikhrninM, Gooch B. Color Transfer between Images[J]. IEEE Computer Graphics and Applications,2001,10:34-40
[7]Welsh T, Mueller K. Transfer color to grey scale images [J]. ACM Transactions on Graphics,2002,21(3):277-280
[8]赵国英,向世明,李华.高阶矩在颜色传输中的应用[J].计算机辅助设计与图形学报,2004,16(1)：62～66
[9]胡国飞,傅健,彭群生.自适应颜色迁移[J].计算机学报,2004,27(9)：1245-1249
[10]Li Hong, Xiang Yao. Colorization of Gray Images Based on Histogram Mapping and Image Layers[J]. Journal of Chinese Computer Systems,2007,28(6):1110～ 1114
[11]Xiang Yao, Zou Beiji, Wang Hui. Multi—source color transfer for natural images[C]. In:The 15th IEEE International Conference on Image Processing, San Diego, CA, USA,2008.469～472
[12]Xiang Yao, Zou Beiji, Li Hong. Selective color transfer with multi～source images[J]. Pattern Recognition Letters,2009,7(30):682-689
[13]贾云涛,胡事民.基于图切分的交互式图像染色算法[J].计算机学报,2006,29(3)：508～512
[14]Rother C, Blake A, Kolmogorov V. Grabcut:interactive foreground extraction using iterated graph cuts[J]. ACM Transactions on Graphics,2004,23(3):309～ 314
[15]Horiuchi T. Colorization algorithm using probabilistic relaxation[J]. Image and Vision Computing,2004,22(3):197～202
[16]Horiuchi T, Hirano S. Colorization algorithm for grayscales image by propagating seed pixels[C]. In:Proceedings of IEEE International Conference on Pattern Recognition,2003,1:457-460
[17]Levin A, Lischinski D, Weiss Y. Colorization using optimization[J]. ACM Transactions on Graphics,2004,23(3):689-694
[18]Yatziv L, Sapiro G. Fast image and video colorization using chrominance blending[J]. IEEE Transactions on Image Processing,2006,15(5):1120-1129
[19]Li ZH, Jing ZL, Yang XH. Color transfer based remote sensing image fusion using non-separable wavelet frame transform[J]. Pattern Recognition Letters, 2005,26(13):2006-2014
[20]张砚,王昱.基于优化的遥感影像彩色化[J].中国科技论文在线,2008,3(2)：114～118
[21]董兆华,张明敏,潘志庚.一种给黑白视频着色的方法[J].计算机工程,2004,30(4)：56～57
[22]崔会卿.基于聚类和距离变换的数字图像彩色化技术[D].石家庄：河北师范大学,2009
[23]MADEIRA JS, STORK A, GROB MH. An approach to computer-supported cartooning [J]. The Visual Computer,1996,12(1):1 -17
[24]Sykora D, Burianek J. Segmentation of black and white cartoons[C]. In: Proceedings of the 19th Spring Conference on Computer Graphics, New York, NY, USA,2003.245-254
[25]赵国英,李华.颜色传输技术在影视制作和处理中的应用[J].影视技术,2003(6)：39-40
[26]Zheng Y F, Edward A E. A local-coloring method for night-vision colorization utilizing image analysis and fusion[J]. Information Fusion,2008,9(2):186-199
[27]向遥.古建筑虚拟色彩修复及漫游技术研究[D].长沙：中南大学信息科学与工程学院,2007
[28]饶娜.基于聚类的无监督图像颜色传递算法研究[D].杭州：浙江大学,2005
[29]刘直芳,王运琼等著.数字图像处理与分析[M].北京：清华大学出版社,2006
[30]滕升华,谌安军,袁建华.一种改进的基于过度分割的图像彩色化方法[J].电子与信息学报,2006,28(7)：1160～1164
[31]滕升华.基于聚类的无监督图像颜色传递算法研究[D].北京：中国科学院研究生院,2006
[32]Tomohisa Takahama, Takahiko Horiuchi, Hiroaki Kotera. Improvement on colorization accuracy by partitioning algorithm in CIELAB color space[J]. Lecture Note in Computer Science,2005,2005(3332):794■801
[33]Levin A, Lischinski D, Weiss Y. Colorization using optimization[J]. ACM Transactions on Graphics,2004,23(3):689-694
[34]Demmel J. Lecture Notes on Numerical Linear Algebra[R]. Berkeley Lecture Notes in Mathematics, Mathematics Dept, UC Berkeley
[35]滕升华,谌安军,邹谋炎.一种基于拉普拉斯方程的图像彩色化方法[J].中国图象图形学报,2006,11(4)：545-548
[36]Caselles V, Coll B, Morel J M. Geometry and color in natural Images [J]. Journal of Mathematical Imaging and Vision,2002,2(16):89-105
[37]MacAdamD. Sources of Color Science[M]. Cambridge:MIT Press,1970
[38]Anany Levitin. Introduction to the Design and Analysis of Algorithms [M]. University:Addison Wesley/Pearson,2003
[39]李志永,滕升华.基于不平度的快速图像彩色化方法[J].中国科学院研究生院学报,2008,25(2)：233-237
[40]滕升华,谌安军.一种改进的基于过度分割的图像彩色化算法[J].电子与信息学报,2006,28(7)：1160-1163
[41]Gomez D, Montero J. A graph coloring approach for image segmentation[J]. Omege,2007,35(2):173 - 183
[42]Nikhil R P, Sankar K P. A review on image segmentation techniques[J]. Pattern Recognition,1993,26:1277-1294
[43]Bergholm F. Edge focusing[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,1987,9(9):726-741
[44]应骏,叶秀清,顾伟康.一个基于知识的边沿提取算法[J].中国图象图形学报,1999,41(3)：239-242
[45]Michael H F Wilkinson. Optimizing Edge Detectors for Robust Automatic Threshold Selection:Coping with Edge Curvature and Noise[J]. Graphical Models and Image Processing,1998,60(5):385～401
[46]Bhabatosh Chanda, Malay K Kundu, Vani Y Padmaja. A Multi-Scale Morphologic Edge Detector[J]. Pattern Recognition,1998,31(10):1469～1478
[47]ZiouD, TabboneS. A multi-scale edge detector[J]. Pattern Recognition,1993, 26(9):1305-1314
[48]Brown M A, Blackwell K T, Khalak H G. Multi-Scale Edge Detection and Feature Binding:An Integrated Approach[J]. Pattern recognition,1998, 31(10):1479～1490
[49]Marr. D著.姚国正译.视觉计算理论[M].北京：科学出版社,1988
[50]黄长专,王彪.图像分割方法研究[J].计算机技术与发展,2009(6)：76～79
[51]罗希平,田捷.图像分割方法综述[J].模式识别与人工智能,1999.3(12)：300～312
[52]Brink A D. Tresholding of digital images using two dimensional entrop-ies[J]. Pattern Recognition,1992,25(8):803～808
[53]杨晖,曲秀杰.图像分割方法综述[J].电脑开发与应用,2005,18(3)：21～23
[54]范静辉,吴建华.基于矢量量化和区域生长的彩色图像分割新算法[J].中国图象图形学报,2005,10(9)：1079～1081
[55]Vincent L, Soille P. Watersheds in digital spaces:an efficient algorithm based on immersion simulations[J]. IEEE Transactions on Pattern Analysis and Machine Intelligent,1991,13(6):583-598
[56]杨家红,刘杰.结合分水岭与自动种子区域生长的彩色图像分割算法[J].中国图象图形学报,2010,1(13)：63～68
[57]范伟.基于区域生长的彩色图像分割算法[J].计算机工程,2010,36(13)：192～194
[58]Cormen TH著.潘金贵等译.算法导论[M].北京：机械工业出版社,2006
[59]Wei Qiyan, Mohan S Kankanhalli. Colorizing infrared home videos[C]. In: IEEE International Conference on Multimedia and Expo, Krakow, Poland, 2003,1:97-100
[60]A. Murat Tekalp.数字视频处理[M].北京：清华大学出版社,1998
[61]Pan Zhigeng, Dong Zhaohua, Zhang Mingmin. A new algorithm for adding color to video or animation clips[C]. In:The 12th International Conference on Computer Graphics, Czech Republic,2004.515～519
[2]Marshall P. Any Color You Like [EB/OL]. Available:http://photography .about.com/library/weekly/aa061002a.htm,2002-07-11
[3]Museum of Broadcast Communication:Encyclopedia of Television [EB/OL]. Available:http://www.museum.tv/archives/etv/C/htmlC/colorization/colorization .htm,2008-06-11
[4]Sykora D, Burianek J, and ZaraJ. Unsupervised colorization of black and white cartoons[C], In:Proceedings of the 3th international symposium on Non-photo-realistic animation and rendering, Annecy, France,2004,7:21-127
[5]Daniel LR, Thomas WC, Chuan CC. Statistics of Cone Responses to Natural Images:Implications for Visual Coding[J]. Journal of the OpticM Society of America,1998,15(8):2036-2045
[6]Reinhard E, AshikhrninM, Gooch B. Color Transfer between Images[J]. IEEE Computer Graphics and Applications,2001,10:34-40
[7]Welsh T, Mueller K. Transfer color to grey scale images [J]. ACM Transactions on Graphics,2002,21(3):277-280
[8]赵国英,向世明,李华.高阶矩在颜色传输中的应用[J].计算机辅助设计与图形学报,2004,16(1)：62～66
[9]胡国飞,傅健,彭群生.自适应颜色迁移[J].计算机学报,2004,27(9)：1245-1249
[10]Li Hong, Xiang Yao. Colorization of Gray Images Based on Histogram Mapping and Image Layers[J]. Journal of Chinese Computer Systems,2007,28(6):1110～ 1114
[11]Xiang Yao, Zou Beiji, Wang Hui. Multi—source color transfer for natural images[C]. In:The 15th IEEE International Conference on Image Processing, San Diego, CA, USA,2008.469～472
[12]Xiang Yao, Zou Beiji, Li Hong. Selective color transfer with multi～source images[J]. Pattern Recognition Letters,2009,7(30):682-689
[13]贾云涛,胡事民.基于图切分的交互式图像染色算法[J].计算机学报,2006,29(3)：508～512
[14]Rother C, Blake A, Kolmogorov V. Grabcut:interactive foreground extraction using iterated graph cuts[J]. ACM Transactions on Graphics,2004,23(3):309～ 314
[15]Horiuchi T. Colorization algorithm using probabilistic relaxation[J]. Image and Vision Computing,2004,22(3):197～202
[16]Horiuchi T, Hirano S. Colorization algorithm for grayscales image by propagating seed pixels[C]. In:Proceedings of IEEE International Conference on Pattern Recognition,2003,1:457-460
[17]Levin A, Lischinski D, Weiss Y. Colorization using optimization[J]. ACM Transactions on Graphics,2004,23(3):689-694
[18]Yatziv L, Sapiro G. Fast image and video colorization using chrominance blending[J]. IEEE Transactions on Image Processing,2006,15(5):1120-1129
[19]Li ZH, Jing ZL, Yang XH. Color transfer based remote sensing image fusion using non-separable wavelet frame transform[J]. Pattern Recognition Letters, 2005,26(13):2006-2014
[20]张砚,王昱.基于优化的遥感影像彩色化[J].中国科技论文在线,2008,3(2)：114～118
[21]董兆华,张明敏,潘志庚.一种给黑白视频着色的方法[J].计算机工程,2004,30(4)：56～57
[22]崔会卿.基于聚类和距离变换的数字图像彩色化技术[D].石家庄：河北师范大学,2009
[23]MADEIRA JS, STORK A, GROB MH. An approach to computer-supported cartooning [J]. The Visual Computer,1996,12(1):1 -17
[24]Sykora D, Burianek J. Segmentation of black and white cartoons[C]. In: Proceedings of the 19th Spring Conference on Computer Graphics, New York, NY, USA,2003.245-254
[25]赵国英,李华.颜色传输技术在影视制作和处理中的应用[J].影视技术,2003(6)：39-40
[26]Zheng Y F, Edward A E. A local-coloring method for night-vision colorization utilizing image analysis and fusion[J]. Information Fusion,2008,9(2):186-199
[27]向遥.古建筑虚拟色彩修复及漫游技术研究[D].长沙：中南大学信息科学与工程学院,2007
[28]饶娜.基于聚类的无监督图像颜色传递算法研究[D].杭州：浙江大学,2005
[29]刘直芳,王运琼等著.数字图像处理与分析[M].北京：清华大学出版社,2006
[30]滕升华,谌安军,袁建华.一种改进的基于过度分割的图像彩色化方法[J].电子与信息学报,2006,28(7)：1160～1164
[31]滕升华.基于聚类的无监督图像颜色传递算法研究[D].北京：中国科学院研究生院,2006
[32]Tomohisa Takahama, Takahiko Horiuchi, Hiroaki Kotera. Improvement on colorization accuracy by partitioning algorithm in CIELAB color space[J]. Lecture Note in Computer Science,2005,2005(3332):794■801
[33]Levin A, Lischinski D, Weiss Y. Colorization using optimization[J]. ACM Transactions on Graphics,2004,23(3):689-694
[34]Demmel J. Lecture Notes on Numerical Linear Algebra[R]. Berkeley Lecture Notes in Mathematics, Mathematics Dept, UC Berkeley
[35]滕升华,谌安军,邹谋炎.一种基于拉普拉斯方程的图像彩色化方法[J].中国图象图形学报,2006,11(4)：545-548
[36]Caselles V, Coll B, Morel J M. Geometry and color in natural Images [J]. Journal of Mathematical Imaging and Vision,2002,2(16):89-105
[37]MacAdamD. Sources of Color Science[M]. Cambridge:MIT Press,1970
[38]Anany Levitin. Introduction to the Design and Analysis of Algorithms [M]. University:Addison Wesley/Pearson,2003
[39]李志永,滕升华.基于不平度的快速图像彩色化方法[J].中国科学院研究生院学报,2008,25(2)：233-237
[40]滕升华,谌安军.一种改进的基于过度分割的图像彩色化算法[J].电子与信息学报,2006,28(7)：1160-1163
[41]Gomez D, Montero J. A graph coloring approach for image segmentation[J]. Omege,2007,35(2):173 - 183
[42]Nikhil R P, Sankar K P. A review on image segmentation techniques[J]. Pattern Recognition,1993,26:1277-1294
[43]Bergholm F. Edge focusing[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,1987,9(9):726-741
[44]应骏,叶秀清,顾伟康.一个基于知识的边沿提取算法[J].中国图象图形学报,1999,41(3)：239-242
[45]Michael H F Wilkinson. Optimizing Edge Detectors for Robust Automatic Threshold Selection:Coping with Edge Curvature and Noise[J]. Graphical Models and Image Processing,1998,60(5):385～401
[46]Bhabatosh Chanda, Malay K Kundu, Vani Y Padmaja. A Multi-Scale Morphologic Edge Detector[J]. Pattern Recognition,1998,31(10):1469～1478
[47]ZiouD, TabboneS. A multi-scale edge detector[J]. Pattern Recognition,1993, 26(9):1305-1314
[48]Brown M A, Blackwell K T, Khalak H G. Multi-Scale Edge Detection and Feature Binding:An Integrated Approach[J]. Pattern recognition,1998, 31(10):1479～1490
[49]Marr. D著.姚国正译.视觉计算理论[M].北京：科学出版社,1988
[50]黄长专,王彪.图像分割方法研究[J].计算机技术与发展,2009(6)：76～79
[51]罗希平,田捷.图像分割方法综述[J].模式识别与人工智能,1999.3(12)：300～312
[52]Brink A D. Tresholding of digital images using two dimensional entrop-ies[J]. Pattern Recognition,1992,25(8):803～808
[53]杨晖,曲秀杰.图像分割方法综述[J].电脑开发与应用,2005,18(3)：21～23
[54]范静辉,吴建华.基于矢量量化和区域生长的彩色图像分割新算法[J].中国图象图形学报,2005,10(9)：1079～1081
[55]Vincent L, Soille P. Watersheds in digital spaces:an efficient algorithm based on immersion simulations[J]. IEEE Transactions on Pattern Analysis and Machine Intelligent,1991,13(6):583-598
[56]杨家红,刘杰.结合分水岭与自动种子区域生长的彩色图像分割算法[J].中国图象图形学报,2010,1(13)：63～68
[57]范伟.基于区域生长的彩色图像分割算法[J].计算机工程,2010,36(13)：192～194
[58]Cormen TH著.潘金贵等译.算法导论[M].北京：机械工业出版社,2006
[59]Wei Qiyan, Mohan S Kankanhalli. Colorizing infrared home videos[C]. In: IEEE International Conference on Multimedia and Expo, Krakow, Poland, 2003,1:97-100
[60]A. Murat Tekalp.数字视频处理[M].北京：清华大学出版社,1998
[61]Pan Zhigeng, Dong Zhaohua, Zhang Mingmin. A new algorithm for adding color to video or animation clips[C]. In:The 12th International Conference on Computer Graphics, Czech Republic,2004.515～519