基于图像类推的人脸画像生成算法与画像检索研究
|
摘要
画像识别是人脸识别的一个全新分支。在刑侦破案或反恐追逃等应用领域,犯罪嫌疑人的照片往往无法直接获取,通常情况下会通过目击者和画家的合作得到其模拟画像,然后在已有的照片数据库中进行检索,进而实现犯罪嫌疑人身份的确认和识别。由于画像与照片的产生机理和信息表达方式不同,两者之间存在较大的几何形变及纹理和灰度差异,若直接采用画像与照片之间的匹配程度进行识别势必产生很大误差。因此,如何减小二者的差异成为画像-照片识别研究的重点和难点,从而画像生成成为人脸画像-照片识别中的关键技术。
本文首先深入研究了图像类推的算法,并在此基础上提出基于图像类推的模拟画像的生成方法,由于传统的图像类推方法在合成模拟画像时在合成时间和合成效果上都不能满足实验的要求,针对上述问题本文对传统的图像类推算法进行了改进,在实验中本文分别从分块类推、参数选择、像素的搜索邻域与像素特征向量的选取等几个方面对传统的图像类推进行改进,使得它能更好的利用于模拟画像的生成,最终实验得到了比较满意的结果。
最后本文采用了基于小波加DCT变换的人脸模拟画像的特征提取方法,它通过对经过小波变换的人脸模拟画像的低频子图进行DCT变换,然后利用最近邻法进行分类检索,在本文中93幅人脸模拟画像数据库中进行实验。实验表明这样的方法能够获得比直接应用DCT变换提取特征更好的识别率。
Sketch recognition is a new branch of face recognition. In applications of case-solving and suspect-searching, to deal with the problem that the photo of a suspect is unavailable in most case, a simulated sketch is generated by cooperation of artists and eyewitnesses as a substitute, so that the content-based image retrieval for identification is conducted in the existing photo database. Due to large differences between a photo-sketch pair caused by different generation mechanism and information expression manner, it is impossible to recognize by simply matching the photo-sketch pair. As a result, how to reduce the difference becomes important for sketch-photo recognition, and sketch synthesis becomes the key technique of recognition.
Firstly the paper studies the algorithms of image analogies and presents a face sketch synthesis method based on image analogies. Because the traditional image analogy method lost too much time and the results of synthesized sketch is not ideal. So a novel image analogy algorithm is proposed in this paper. To improve the synthesis performance, some novel idea is introduced on sub-block dividing image analogy, chosen of image analogy parameters, neighborhood of the pixel, and the eigenvector of the pixel in the experiment. Experimental results show that the novel algorithm can achieve good results in face sketch synthesis.
Finally, a face sketch recognition method based on wavelet transform and discrete c osine transform is presented. First the original image was decomposed by wavelet transf orm. Then, the DCT method can be directly used to the low frequent sub-image obtain ed form on the previous step. In the end, the classical K-Nearest Neighbor method is p erformed on the reductions for the pattern classification. Experiments over a data set co ntaining 93 people show that the above mentioned algorithm performs higher accuracy o f recognition than discrete cosine transform method.
本文首先深入研究了图像类推的算法,并在此基础上提出基于图像类推的模拟画像的生成方法,由于传统的图像类推方法在合成模拟画像时在合成时间和合成效果上都不能满足实验的要求,针对上述问题本文对传统的图像类推算法进行了改进,在实验中本文分别从分块类推、参数选择、像素的搜索邻域与像素特征向量的选取等几个方面对传统的图像类推进行改进,使得它能更好的利用于模拟画像的生成,最终实验得到了比较满意的结果。
最后本文采用了基于小波加DCT变换的人脸模拟画像的特征提取方法,它通过对经过小波变换的人脸模拟画像的低频子图进行DCT变换,然后利用最近邻法进行分类检索,在本文中93幅人脸模拟画像数据库中进行实验。实验表明这样的方法能够获得比直接应用DCT变换提取特征更好的识别率。
Sketch recognition is a new branch of face recognition. In applications of case-solving and suspect-searching, to deal with the problem that the photo of a suspect is unavailable in most case, a simulated sketch is generated by cooperation of artists and eyewitnesses as a substitute, so that the content-based image retrieval for identification is conducted in the existing photo database. Due to large differences between a photo-sketch pair caused by different generation mechanism and information expression manner, it is impossible to recognize by simply matching the photo-sketch pair. As a result, how to reduce the difference becomes important for sketch-photo recognition, and sketch synthesis becomes the key technique of recognition.
Firstly the paper studies the algorithms of image analogies and presents a face sketch synthesis method based on image analogies. Because the traditional image analogy method lost too much time and the results of synthesized sketch is not ideal. So a novel image analogy algorithm is proposed in this paper. To improve the synthesis performance, some novel idea is introduced on sub-block dividing image analogy, chosen of image analogy parameters, neighborhood of the pixel, and the eigenvector of the pixel in the experiment. Experimental results show that the novel algorithm can achieve good results in face sketch synthesis.
Finally, a face sketch recognition method based on wavelet transform and discrete c osine transform is presented. First the original image was decomposed by wavelet transf orm. Then, the DCT method can be directly used to the low frequent sub-image obtain ed form on the previous step. In the end, the classical K-Nearest Neighbor method is p erformed on the reductions for the pattern classification. Experiments over a data set co ntaining 93 people show that the above mentioned algorithm performs higher accuracy o f recognition than discrete cosine transform method.
引文
[1]何东风,凌截.人脸识别技术综述[J].微机发展,2003,13(12)：75-78
[2]Brunelli R,Poggio T.Face recognition:features versus template[J].IEEE Trans Pattern Analysis and Machine Intelligence,1993,15(10):1042-1052
[3]Yuille A L.Deformable templates for face recognition[J].Journal of Cognitive Neuroscience,1 991,3(1):59-70
[4]Kass M,Witkin A,Terzopoulos D.Snakes:active contour models[J].International Journal of Computer Vision,1988,1(6):321-331
[5]Hunag C L,Chen C W.Human.facial feature extraction for face interpretation and recognition [J].Pattern Recognition,1992,25(12):1435-1444
[6]Juanjuan Zhong,Xinbo Gao,Chunna Tian.Face Sketch Synthesis Using E-HMM and Selective Ensemble[J].IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP),2007,15(8):134-142
[7]Turk M,Pentland A.Eigenfaces for recognition[J].Cognitive Neuroscience,1991,3(1):71-86
[8]Turk M,Pentland A.Face Recognition using Eigenfaces[C].IEEE Computer Society Confernce on Computer Vision and Pattern Recognition,1991:586-591
[9]H.Peng,D.Dual EIgenspace Method for Human Face recognition[J].Electronics Lettres,1997,33 (5):283-284
[10]Hong Z Q.Algebraic feature extraction of image for recognition[J].Pattern Recognition,1991, 24(3):211-219
[11]M.Lades,J.C.Vorbriiggen,J.Buhmann,J.Lange,C.Malsburg,R.P.Wurtz.,W.Konen.Distortion invariant Recognition in the dynamic link architecture[J].IEEE Trans.on Computers,1993,3(42):300-311
[12]王磊,莫玉龙,戚飞虎.眼球的自动定位[J].红外与毫米波学报,1998,17(5)：349-356
[13]Rao. K,Yip. P. Discrete Cosine Transform-Algorithms[J]. Advantages, Applications. NewYork, Academic.1990
[14]Rowley H A,Baluja S,Kanade T.Neural network-based face detection[J].IEEE Trans.Pattern Analysis and Machine Intelligence,1998,20 (1):23-38
[15]马利克.人脸模拟画像检索研究[D].西安：西北工业大学,2007
[16]Rhodes G. Secrets of the face[J]. New Zealand Journal of Psychology,1994,23(1):3-17
[17]Li Y, Kobatake H.Extraction of facial sketch based on morphological processing[C].In:Proceedings of the 1997 IEEE International Conference on Image Processing Barbara, CA, USA,1997:316-319
[18]陈洪,郑南宁,徐迎庆,沈向洋.基于样本学习的人像线条画生成系统[J].软件学报,2003,14(2)：202-208
[19]阎芳,费广正,柳婷婷,马文慧,石民勇.漫画风格的人脸肖像生成算法[J].计算机辅助设计与图形学学报,2007,19(4)：442-447
[20]陈洪,郑南宁,梁林,徐迎庆,沈向洋.基于样本学习的肖像画自动生成算法[J].计算机学报,2003,26(2)：147-152
[21]Wang J,Bao H J,Zhou W H. Automatic image-based pencil sketch rendering[J].Journal of Computer Science and Technology,2002,17(5):347-356
[22]Tang X O,Wang X GFace photo recognition using sketch[C].In:Proceedings of International Conference on Image Processing, Rochester, New York,2002:257-260
[23]Tang X O,Wang X GFace sketch synthesis and recognition[C].In:Proceedings of International Conference on Computer Vision,Nice,France,2003:687-694
[24]丁宾.人脸画像识别算法研究[D].西安：西安电子科技大学,2004
[25]张英武.基于画像的人脸识别方法研究[D].西安：西安电子科技大学,2005
[26]刘峰,刘志镜.手绘人脸识别的研究[J].现代电子技术,2006,12(10)：41-43
[27]Hertzmann A,Jacobs Charles E,Nuria Oliver,et al.Image Analogies[C].Proceedings of SIGGR APH2001,2001:327-340
[28]Michael Ashikhmin.Synthesizing Natural Textures[C].In:2001 ACM Symposium on Interactive 3D Graphics,2001:217-226
[29]钟娟娟.基于机器学习的人脸画像自动生成算法研究[D].西安：西安电子科技大学,2007
[30]丁宾,高新波,姬红兵.基于离散余弦变换的人脸画像识别方法[J].计算机工程,2004,30(20)：151-153
[3l]贾永生.浅议模拟画像在侦查中的应用[J].政法学刊.2004,21(3)：76-79
[32]张翠平,苏光大.人脸识别技术综述[J].中国图象图形学报,2000,5(11)：885-894
[33]姚玉荣,章毓晋.利用小波和矩进行基于形状的图象检索[J].中国图象图形学报,2000,5(3)：206-210
[34]吴明珠.利用成对几何直方图改进基于轮廓的图像检索[J].多媒体技术及其应用,2007,16(7),1140-1147
[35]樊亚春,耿国华,周明全.用不变矩和边界方向进行形状检索[J].小型微型计算机系统,2004,25(4)：659-662
[36]谢颖,张雪飞.一种新的主动轮廓线模型[J].中国科技信息,2007,21(16)：276-277
[37]刘亘铁.基于内容的图像检索关键技术研究与系统设计[D].西安：西安电子科技大学,2006
[38]丁险峰,吴洪,张宏江,马颂德.形状匹配综述[J].自动化学报,2001,27(5)：678-693
[39]程博,李育才,吴国平.一种改进的基于形状的图像检索方法[J].软件导刊,2005,10(16)：26-27
[40]顾华,苏光大,杜成.人脸识别中的眼角定位和预处理[J].中国图象图形学报,2002,11(3)439-440
[41]X.Tang,X.Wang. Face Sketch Recognition[J].Image Processing,2004,14(1):50-57
[42]Chen H,Zheng N N,Xu Y Q, Shum H Y.An exampled-based facial sketch.generation system[J]. Journal of Software,2003,14(2):202-208
[43]Liu Q S, Tang X O.A nonlinear approach for face sketch synthesis and recognition[C].In: Proceedings of International Conference on CVPR, San Diego,CA,USA,2005:1005-1010
[44]童晓阳,杨名利.人脸轮廓定位和归一化的研究[J].系统仿真学报,2005,17(3)：701-709
[45]姚春生.基于目标轮廓的图像检索模型[D].广州：暨南大学,2005
[46]马丽丽.基于PSO快速图像类推及其应用[D].大连：大连理工大学,2006
[47]http://people.csail.mit.edu/xgwang/sketch_multiscale.html
[2]Brunelli R,Poggio T.Face recognition:features versus template[J].IEEE Trans Pattern Analysis and Machine Intelligence,1993,15(10):1042-1052
[3]Yuille A L.Deformable templates for face recognition[J].Journal of Cognitive Neuroscience,1 991,3(1):59-70
[4]Kass M,Witkin A,Terzopoulos D.Snakes:active contour models[J].International Journal of Computer Vision,1988,1(6):321-331
[5]Hunag C L,Chen C W.Human.facial feature extraction for face interpretation and recognition [J].Pattern Recognition,1992,25(12):1435-1444
[6]Juanjuan Zhong,Xinbo Gao,Chunna Tian.Face Sketch Synthesis Using E-HMM and Selective Ensemble[J].IEEE International Conference on Acoustics Speech and Signal Processing (ICASSP),2007,15(8):134-142
[7]Turk M,Pentland A.Eigenfaces for recognition[J].Cognitive Neuroscience,1991,3(1):71-86
[8]Turk M,Pentland A.Face Recognition using Eigenfaces[C].IEEE Computer Society Confernce on Computer Vision and Pattern Recognition,1991:586-591
[9]H.Peng,D.Dual EIgenspace Method for Human Face recognition[J].Electronics Lettres,1997,33 (5):283-284
[10]Hong Z Q.Algebraic feature extraction of image for recognition[J].Pattern Recognition,1991, 24(3):211-219
[11]M.Lades,J.C.Vorbriiggen,J.Buhmann,J.Lange,C.Malsburg,R.P.Wurtz.,W.Konen.Distortion invariant Recognition in the dynamic link architecture[J].IEEE Trans.on Computers,1993,3(42):300-311
[12]王磊,莫玉龙,戚飞虎.眼球的自动定位[J].红外与毫米波学报,1998,17(5)：349-356
[13]Rao. K,Yip. P. Discrete Cosine Transform-Algorithms[J]. Advantages, Applications. NewYork, Academic.1990
[14]Rowley H A,Baluja S,Kanade T.Neural network-based face detection[J].IEEE Trans.Pattern Analysis and Machine Intelligence,1998,20 (1):23-38
[15]马利克.人脸模拟画像检索研究[D].西安：西北工业大学,2007
[16]Rhodes G. Secrets of the face[J]. New Zealand Journal of Psychology,1994,23(1):3-17
[17]Li Y, Kobatake H.Extraction of facial sketch based on morphological processing[C].In:Proceedings of the 1997 IEEE International Conference on Image Processing Barbara, CA, USA,1997:316-319
[18]陈洪,郑南宁,徐迎庆,沈向洋.基于样本学习的人像线条画生成系统[J].软件学报,2003,14(2)：202-208
[19]阎芳,费广正,柳婷婷,马文慧,石民勇.漫画风格的人脸肖像生成算法[J].计算机辅助设计与图形学学报,2007,19(4)：442-447
[20]陈洪,郑南宁,梁林,徐迎庆,沈向洋.基于样本学习的肖像画自动生成算法[J].计算机学报,2003,26(2)：147-152
[21]Wang J,Bao H J,Zhou W H. Automatic image-based pencil sketch rendering[J].Journal of Computer Science and Technology,2002,17(5):347-356
[22]Tang X O,Wang X GFace photo recognition using sketch[C].In:Proceedings of International Conference on Image Processing, Rochester, New York,2002:257-260
[23]Tang X O,Wang X GFace sketch synthesis and recognition[C].In:Proceedings of International Conference on Computer Vision,Nice,France,2003:687-694
[24]丁宾.人脸画像识别算法研究[D].西安：西安电子科技大学,2004
[25]张英武.基于画像的人脸识别方法研究[D].西安：西安电子科技大学,2005
[26]刘峰,刘志镜.手绘人脸识别的研究[J].现代电子技术,2006,12(10)：41-43
[27]Hertzmann A,Jacobs Charles E,Nuria Oliver,et al.Image Analogies[C].Proceedings of SIGGR APH2001,2001:327-340
[28]Michael Ashikhmin.Synthesizing Natural Textures[C].In:2001 ACM Symposium on Interactive 3D Graphics,2001:217-226
[29]钟娟娟.基于机器学习的人脸画像自动生成算法研究[D].西安：西安电子科技大学,2007
[30]丁宾,高新波,姬红兵.基于离散余弦变换的人脸画像识别方法[J].计算机工程,2004,30(20)：151-153
[3l]贾永生.浅议模拟画像在侦查中的应用[J].政法学刊.2004,21(3)：76-79
[32]张翠平,苏光大.人脸识别技术综述[J].中国图象图形学报,2000,5(11)：885-894
[33]姚玉荣,章毓晋.利用小波和矩进行基于形状的图象检索[J].中国图象图形学报,2000,5(3)：206-210
[34]吴明珠.利用成对几何直方图改进基于轮廓的图像检索[J].多媒体技术及其应用,2007,16(7),1140-1147
[35]樊亚春,耿国华,周明全.用不变矩和边界方向进行形状检索[J].小型微型计算机系统,2004,25(4)：659-662
[36]谢颖,张雪飞.一种新的主动轮廓线模型[J].中国科技信息,2007,21(16)：276-277
[37]刘亘铁.基于内容的图像检索关键技术研究与系统设计[D].西安：西安电子科技大学,2006
[38]丁险峰,吴洪,张宏江,马颂德.形状匹配综述[J].自动化学报,2001,27(5)：678-693
[39]程博,李育才,吴国平.一种改进的基于形状的图像检索方法[J].软件导刊,2005,10(16)：26-27
[40]顾华,苏光大,杜成.人脸识别中的眼角定位和预处理[J].中国图象图形学报,2002,11(3)439-440
[41]X.Tang,X.Wang. Face Sketch Recognition[J].Image Processing,2004,14(1):50-57
[42]Chen H,Zheng N N,Xu Y Q, Shum H Y.An exampled-based facial sketch.generation system[J]. Journal of Software,2003,14(2):202-208
[43]Liu Q S, Tang X O.A nonlinear approach for face sketch synthesis and recognition[C].In: Proceedings of International Conference on CVPR, San Diego,CA,USA,2005:1005-1010
[44]童晓阳,杨名利.人脸轮廓定位和归一化的研究[J].系统仿真学报,2005,17(3)：701-709
[45]姚春生.基于目标轮廓的图像检索模型[D].广州：暨南大学,2005
[46]马丽丽.基于PSO快速图像类推及其应用[D].大连：大连理工大学,2006
[47]http://people.csail.mit.edu/xgwang/sketch_multiscale.html