人脸老化模拟与PCA算法结合的人脸识别模型
|
摘要
针对如何在读取两张年龄有较大跨度的人物面部照片的情况下来进行人脸识别这个问题,该文给出了较有创新性的方法。首先使用高斯模型,肤色信息人脸检测的分割算法,迭代算法提取人脸特征部位。然后使用基于人脸外轮廓的局部曲率标准差的个性化脸型匹配算法结合关于形状和颜色的衰老合成模型和年龄原型的纹理增强合成模型给出了基于个性化原型的人脸老化图像合成方法。接着在传统PCA算法的基础上给出了结合人脸老化模拟的改进型算法思路,从人脸不易变量和易变量的重建两方面入手判定人脸的相似性,理论上可以较为准确地在跨年龄情况下进行人脸识别。
In order to solve the problem of how to read two large-span facial photos of people,this paper presents a more innovative method.Firstly,we use Gauss model,skin color information face detection segmentation al-gorithm,and iterative algorithm to extract face feature parts.Then,a personalized face pattern matching algori-thm based on local curvature standard deviation of face contour is used to synthesize the aging image based on personalized prototype,which combines the aging synthesis model of shape and color and the texture en-hancement synthesis model of age prototype.Then,on the basis of the traditional PCA algorithm,an improved algorithm combined with face aging simulation is proposed.The similarity of face can be judged from two aspects: the reconstruction of face variables and the reconstruction of face variables.In theory,face recognition can be carried out more accurately in cross-age situations.
In order to solve the problem of how to read two large-span facial photos of people,this paper presents a more innovative method.Firstly,we use Gauss model,skin color information face detection segmentation al-gorithm,and iterative algorithm to extract face feature parts.Then,a personalized face pattern matching algori-thm based on local curvature standard deviation of face contour is used to synthesize the aging image based on personalized prototype,which combines the aging synthesis model of shape and color and the texture en-hancement synthesis model of age prototype.Then,on the basis of the traditional PCA algorithm,an improved algorithm combined with face aging simulation is proposed.The similarity of face can be judged from two aspects: the reconstruction of face variables and the reconstruction of face variables.In theory,face recognition can be carried out more accurately in cross-age situations.
引文
[1]张翀.基于肤色分割和匹配的人脸识别算法研究[D].南昌航空大学科技学院,2013.
[2]高冰钰.基于最大类间方差法图像分割程序设计-迭代阈值法[D].沈阳:沈阳理工大学,2016.
[3]洪浩,霍春宝.基于最大类间方差法与GA相结合的图像分割算法研究[J].辽宁工业大学学报(自然科学版),2016,36(2):99-102.
[4]王章野,曹玫璇,李理,等.基于个性化原型的人脸衰老图像合成[J].电子学报,2009,37(z1):118-124.
[5]Rowland D A,Perrett D I.Manipulating facial appearance through shape and color[J].IEEE Computer Graphics and Applications,1995,15(5):70-76.
[6]脸型也会决定衰老方式?[EB/OL].http://www.mr-fww.com/rj_view.asp?id=414.2007-06-07.Can the shape of the face determine the method of aging?[EB/OL].http://www.mrfww.com/rj_view.asp?id=414.2007-06-07.(in Chinese).
[7]Goshtasby A.Piecewise linear mapping functions for image registration[J].Pattern Recognition:The Journal of the Pattern Recognition Society,1986,19(6):459-466.
[8]Ruprecht D,Muller H.Image warping with scattered data interpolation[J].IEEE Computer Graphics and Applications,1995,15(2):37-43.
[2]高冰钰.基于最大类间方差法图像分割程序设计-迭代阈值法[D].沈阳:沈阳理工大学,2016.
[3]洪浩,霍春宝.基于最大类间方差法与GA相结合的图像分割算法研究[J].辽宁工业大学学报(自然科学版),2016,36(2):99-102.
[4]王章野,曹玫璇,李理,等.基于个性化原型的人脸衰老图像合成[J].电子学报,2009,37(z1):118-124.
[5]Rowland D A,Perrett D I.Manipulating facial appearance through shape and color[J].IEEE Computer Graphics and Applications,1995,15(5):70-76.
[6]脸型也会决定衰老方式?[EB/OL].http://www.mr-fww.com/rj_view.asp?id=414.2007-06-07.Can the shape of the face determine the method of aging?[EB/OL].http://www.mrfww.com/rj_view.asp?id=414.2007-06-07.(in Chinese).
[7]Goshtasby A.Piecewise linear mapping functions for image registration[J].Pattern Recognition:The Journal of the Pattern Recognition Society,1986,19(6):459-466.
[8]Ruprecht D,Muller H.Image warping with scattered data interpolation[J].IEEE Computer Graphics and Applications,1995,15(2):37-43.