用于面诊的嘴巴定位算法研究
|
摘要
针对嘴巴区域特征的精确定位对于中医面诊客观化的研究具有重要意义,提出一种基于双肤色模型、椭圆拟合与改进C-V水平集模型相结合的分割嘴巴的方法。考虑到肤色在像素空间邻域与灰度值域的平滑相似性,首先给出一种二维伽马函数的自适应光照补偿法,提升非均匀光照下肤色聚类的稳定性,进而利用根据实验确定的双肤色模型,进行肤色检测,并采用数学形态学方法去除噪声等影响,Sobel方法提取轮廓;然后根据所得的边缘和初步轮廓利用直接最小二乘椭圆拟合法提取人脸区域;最后采用改进C-V水平集模型对嘴部区域进行分割。实验结果表明,采用该算法能够得到更好的分割效果,满足中医面诊图像分割要求,为面部五官的进一步分割和检测奠定基础。
The precise location according to the characteristics of the mouth area is of great significance for the study of the objectification of traditional Chinese medicine. Therefore, a method based on double skin color model, ellipse fitting and improved C-V level set model is proposed in this study. With the consideration of the smooth similarity of skin color in spatial neighborhood and gray-level domain, an adaptive illumination compensation method based on two-dimensional gamma function is firstly proposed to improve the stability of skin color clustering under non-uniform illumination. Subsequently, the experimentally determined double skin color model is used to perform skin color detection; and mathematical morphology is used to remove noise and other effects; and Sobel method was used to extract the contours. Based on the obtained edge and preliminary contours,the face region was extracted with direct least squares ellipse fitting method. Finally, the mouth area was segmented with the improved C-V level set model. The experimental results reveal that the proposed algorithm can obtain better segmentation results and meet the requirements of facial diagnosis in traditional Chinese medicine, laying a foundation for further segmentation and detection of facial features.
The precise location according to the characteristics of the mouth area is of great significance for the study of the objectification of traditional Chinese medicine. Therefore, a method based on double skin color model, ellipse fitting and improved C-V level set model is proposed in this study. With the consideration of the smooth similarity of skin color in spatial neighborhood and gray-level domain, an adaptive illumination compensation method based on two-dimensional gamma function is firstly proposed to improve the stability of skin color clustering under non-uniform illumination. Subsequently, the experimentally determined double skin color model is used to perform skin color detection; and mathematical morphology is used to remove noise and other effects; and Sobel method was used to extract the contours. Based on the obtained edge and preliminary contours,the face region was extracted with direct least squares ellipse fitting method. Finally, the mouth area was segmented with the improved C-V level set model. The experimental results reveal that the proposed algorithm can obtain better segmentation results and meet the requirements of facial diagnosis in traditional Chinese medicine, laying a foundation for further segmentation and detection of facial features.
引文
[1]张红凯,李福凤,郑晓燕,等.中医面诊图谱颜色、光泽、唇色参数范围研究[C].世界中医药学会联合会, 2015:186-189.ZHANG H K, LI F F, ZHENG X Y, et al. Research on parameters of facial complexion, gloss and lip color in CM[C]. World Federation of Chinese Medicine Societies, 2015:186-189.
[2]郑莉丽,李晓强,李福凤,等.基于支持向量机的中医望诊唇色自动分类[J].生物医学工程杂志, 2011, 28(1):7-11.ZHENG L L, LI X Q, LI F F, et al. Automatic classification of lip color based on SVM in traditional Chinese medicine inspection[J]. Journal of Biomedical Engineering, 2011, 28(1):7-11.
[3]杨云聪,张菁,卓力,等.应用于中医面诊的人脸区域分割方法[J].测控技术, 2012, 31(5):25-28.YANG Y C, ZHANG J, ZHUO L, et al. Face region segmentation applying in face diagnosis of traditional Chinese medicine[J].Measurement&Control Technology, 2012, 31(5):25-28.
[4]朴鑫.面诊中的人脸分割研究[D].哈尔滨:哈尔滨工业大学, 2006.PIAO X. Face segmentation research in face consultation[D]. Harbin:Harbin Institute of Technology, 2006.
[5]宁娟,朱敏,戴李君.基于Ada Boost算法和色彩信息的脸部特征定位[J].计算机应用与软件, 2016, 33(5):207-211.NING J, ZHU M, DAI L J. Localisation of facial features based on Ada Boost algorithm and color information[J]. Computer Applications and Software, 2016, 33(5):207-211.
[6]毛红朝.面向中医面诊的诊断信息提取-关键算法研究与实现[D].厦门:厦门大学, 2007MAO H C. Extraction of diagnostic information for TCM face-to-face diagnosis-research and implementation of key algorithms[D].Xiamen:Xiamen University, 2007.
[7]廉文娟.人脸识别中的特征提取技术[D].青岛:山东科技大学,2002.LIAN W J. Feature extraction technology in face recognition[D].Qingdao:Shandong University of Science and Technology, 2002.
[8]刘志成,王殿伟,刘颖,等.基于二维伽马函数的光照不均匀图像自适应校正算法[J].北京理工大学学报, 2016, 36(2):191-196.LIU Z C, WANG D W, LIU Y, et al. Adaptive adjustment algorithm for non-uniform illumination images based on 2D gamma function[J].Transactions of Beijing Institute of Technology, 2016, 36(2):191-196.
[9]赵艳妮,郭华磊.基于多色彩空间的肤色检测研究[J].计算机技术与发展, 2013, 23(10):250-252.ZHAO Y N, GUO H L. Research of skin detection based on multicolor spaces[J]. Computer Technology and Development, 2013, 23(10):250-252.
[10]BHUIYUAN M A, ALSAADE F W. Genetic search for face detection[C]. Proceedings of the World Congress on Engineering, 2015.
[11]RAFAEL C G, RICHARD E W.数字图像处理[M].第3版.北京:电子工业出版社, 2013.RAFAEL C G, RICHARD E W. Digital image processing using MATLAB[M]. 3rd ed. Beijing:Electronic Industry Press, 2013.
[12]吴绿芳,雷蕴奇,魏昇,等.基于肤色模型和梯度算子的正面人脸轮廓提取[J].电脑与信息技术, 2008, 16(1):1-4.WU L F, LEI Y Q, WEI S, et al. Front face contour extraction based on skin color model and gradient operator[J]. Computer and Information Technology, 2008, 16(1):1-4.
[13]FITZGIBBON A, PILU M, FISHER R B. Direct least square fitting of ellipse[J]. IEEE Trans Pattern Anal Mach Intell, 1999, 21(5):476-480.
[14]金俊才,童卫青,梁晓妮,等.基于椭圆拟合的近红外图像的眼睛精确定位法[J].华东师范大学学报(自然科学版), 2012, 23(3):103-110.JIN J C, TONG W Q, LIANG X N, et al. Accurate eye location in nearinfrared images based on ellipse fitting[J]. Journal of East China Normal University(Natural Science), 2012, 23(3):103-110.
[15]CHAN T F, VESE LA. Active contours without edges[J]. IEEE Trans Image Process, 2001, 10(2):266-277.
[16]MUMFORD D, SHAH J. Optimal approximation by piecewise smooth functions and associated variational problems[J]. Comm Pure Appl Math, 1989, 42(5):577-685.
[17]LI C M, HUANG R, DING Z H, et al. A level set method for image segmentation in the presence of intensity inhomogeneities with application to MRI[J]. IEEE Trans Image Process, 2011, 20(7):2007-2016.
[18]CHAUDHURY K N, RAMAKRISHNAN K R. Stability and convergence of the level set method in computer vision[J]. Pattern Recognit Lett, 2007, 28:884-893.
[2]郑莉丽,李晓强,李福凤,等.基于支持向量机的中医望诊唇色自动分类[J].生物医学工程杂志, 2011, 28(1):7-11.ZHENG L L, LI X Q, LI F F, et al. Automatic classification of lip color based on SVM in traditional Chinese medicine inspection[J]. Journal of Biomedical Engineering, 2011, 28(1):7-11.
[3]杨云聪,张菁,卓力,等.应用于中医面诊的人脸区域分割方法[J].测控技术, 2012, 31(5):25-28.YANG Y C, ZHANG J, ZHUO L, et al. Face region segmentation applying in face diagnosis of traditional Chinese medicine[J].Measurement&Control Technology, 2012, 31(5):25-28.
[4]朴鑫.面诊中的人脸分割研究[D].哈尔滨:哈尔滨工业大学, 2006.PIAO X. Face segmentation research in face consultation[D]. Harbin:Harbin Institute of Technology, 2006.
[5]宁娟,朱敏,戴李君.基于Ada Boost算法和色彩信息的脸部特征定位[J].计算机应用与软件, 2016, 33(5):207-211.NING J, ZHU M, DAI L J. Localisation of facial features based on Ada Boost algorithm and color information[J]. Computer Applications and Software, 2016, 33(5):207-211.
[6]毛红朝.面向中医面诊的诊断信息提取-关键算法研究与实现[D].厦门:厦门大学, 2007MAO H C. Extraction of diagnostic information for TCM face-to-face diagnosis-research and implementation of key algorithms[D].Xiamen:Xiamen University, 2007.
[7]廉文娟.人脸识别中的特征提取技术[D].青岛:山东科技大学,2002.LIAN W J. Feature extraction technology in face recognition[D].Qingdao:Shandong University of Science and Technology, 2002.
[8]刘志成,王殿伟,刘颖,等.基于二维伽马函数的光照不均匀图像自适应校正算法[J].北京理工大学学报, 2016, 36(2):191-196.LIU Z C, WANG D W, LIU Y, et al. Adaptive adjustment algorithm for non-uniform illumination images based on 2D gamma function[J].Transactions of Beijing Institute of Technology, 2016, 36(2):191-196.
[9]赵艳妮,郭华磊.基于多色彩空间的肤色检测研究[J].计算机技术与发展, 2013, 23(10):250-252.ZHAO Y N, GUO H L. Research of skin detection based on multicolor spaces[J]. Computer Technology and Development, 2013, 23(10):250-252.
[10]BHUIYUAN M A, ALSAADE F W. Genetic search for face detection[C]. Proceedings of the World Congress on Engineering, 2015.
[11]RAFAEL C G, RICHARD E W.数字图像处理[M].第3版.北京:电子工业出版社, 2013.RAFAEL C G, RICHARD E W. Digital image processing using MATLAB[M]. 3rd ed. Beijing:Electronic Industry Press, 2013.
[12]吴绿芳,雷蕴奇,魏昇,等.基于肤色模型和梯度算子的正面人脸轮廓提取[J].电脑与信息技术, 2008, 16(1):1-4.WU L F, LEI Y Q, WEI S, et al. Front face contour extraction based on skin color model and gradient operator[J]. Computer and Information Technology, 2008, 16(1):1-4.
[13]FITZGIBBON A, PILU M, FISHER R B. Direct least square fitting of ellipse[J]. IEEE Trans Pattern Anal Mach Intell, 1999, 21(5):476-480.
[14]金俊才,童卫青,梁晓妮,等.基于椭圆拟合的近红外图像的眼睛精确定位法[J].华东师范大学学报(自然科学版), 2012, 23(3):103-110.JIN J C, TONG W Q, LIANG X N, et al. Accurate eye location in nearinfrared images based on ellipse fitting[J]. Journal of East China Normal University(Natural Science), 2012, 23(3):103-110.
[15]CHAN T F, VESE LA. Active contours without edges[J]. IEEE Trans Image Process, 2001, 10(2):266-277.
[16]MUMFORD D, SHAH J. Optimal approximation by piecewise smooth functions and associated variational problems[J]. Comm Pure Appl Math, 1989, 42(5):577-685.
[17]LI C M, HUANG R, DING Z H, et al. A level set method for image segmentation in the presence of intensity inhomogeneities with application to MRI[J]. IEEE Trans Image Process, 2011, 20(7):2007-2016.
[18]CHAUDHURY K N, RAMAKRISHNAN K R. Stability and convergence of the level set method in computer vision[J]. Pattern Recognit Lett, 2007, 28:884-893.
© 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 |