基于Isomap特征降维的人脸表情相似度评估方法
|
摘要
面部表情对影视表演有着紧要的作用,表情相似度度量是评判面部表情是否表演到位的一种方式。现有的表情相似度度量方法存在微小表情变化度量不精确、局部细节描述能力弱等问题。因此,本文在Isomap (Isometric Feature Mapping)特征降维基础上提出了一种新的表情相似度评估方法,将其应用到了表演艺术课中。首先,使用基于梯度提高学习的回归树方法实现面部特征点定位以及追踪,得到特征点的运动信息。然后,将得到的高维特征点运动信息运用非线性降维算法Isomap降维,得到人脸表情曲线。最后,提出了基于时序的相似度匹配方法,通过对比表情序列中每一帧的强度得出表情相似度。实验结果表明,该方法是可行的,能够有效地评估表演者对影视片段中表情模仿的相似度。
引文
[1]李月龙,靳彦,汪剑鸣.人脸特征点提取方法综述[J].计算机学报,2016,39(07)∶1356-1374.
[2]胡敏,江河,王晓华.基于几何和纹理特征的表情层级分类方法[J].电子学报,2017,45(1)∶164-172.
[3]WANG N N,GAO X B,TAO D H.Facial feature point detection∶A comprehensive survey[J].Neurocomputing,2018,275∶50-65.
[4]KARAN R,KAZUYUKI S,HITOSHI I.Occlusion handling in feature point tracking using ranked parts based models[C]//IEEE International Conference on Image Processing.IEEE,2015∶740-744.
[5]黄铉.特征降维技术的研究与进展[J].计算机科学,2018,45(S1)∶16-21+53.
[6]YU W B,ZHANG M,YI S.Learning a Stable Local Manifold Representation for Hyperspe-Ctral Linear Dimensionality Reduction[C]//IEEE International Geoscience and Remote Sensing Symposium IEEE International Geoscience and Remote Sensing Symposium.IEEE,2018∶3555-3558.
[7]WEINBERGER K Q,SAUL L K.Unsupervi-Sed Learning of Image Manifolds by Semidefinite Programming[J].International Journal of Computer Vision,2006,70(1)∶77-90.
[8]LI Y Y.Curvature-aware manifold learning[J].Pattern Recognition,2018,83∶273-286.
[9]TENENBAUM J B,SILVA V D,Langford J C.A Global Geometric Framework for Nonlinear Dimensionality Reduction[J].Science,2000,290∶2319-2323.
[10]邹朋成,王建东,杨国庆.辅助信息自动生成的时间序列距离度量学习[J].软件学报,2013,24(11)∶2642-2655.
[11]Márcio Macedo,Antonio Apolinário.Improved antialiasing for Euclidean distance transform shadow mapping[J].Computers&Graphics,2018,71∶166-179.
[12]Tang J R,Cheng H,Zhao Y.Structured dynamic time warping for continuous hand trajectory gesture recognition[J].Pattern Recognition,2018,80∶21-31.
[13]李正欣,张凤鸣,李克武.一种支持DTW距离的多元时间序列索引结构[J].软件学报,2014,25(3)∶560-575.
[14]CHANG Y,HU C,TURK M.Manifold of Facial Expression[C]//IEEE International Workshop on Analysis and Modeling of Faces and Gestures.IEEE,2003∶28-35.
[2]胡敏,江河,王晓华.基于几何和纹理特征的表情层级分类方法[J].电子学报,2017,45(1)∶164-172.
[3]WANG N N,GAO X B,TAO D H.Facial feature point detection∶A comprehensive survey[J].Neurocomputing,2018,275∶50-65.
[4]KARAN R,KAZUYUKI S,HITOSHI I.Occlusion handling in feature point tracking using ranked parts based models[C]//IEEE International Conference on Image Processing.IEEE,2015∶740-744.
[5]黄铉.特征降维技术的研究与进展[J].计算机科学,2018,45(S1)∶16-21+53.
[6]YU W B,ZHANG M,YI S.Learning a Stable Local Manifold Representation for Hyperspe-Ctral Linear Dimensionality Reduction[C]//IEEE International Geoscience and Remote Sensing Symposium IEEE International Geoscience and Remote Sensing Symposium.IEEE,2018∶3555-3558.
[7]WEINBERGER K Q,SAUL L K.Unsupervi-Sed Learning of Image Manifolds by Semidefinite Programming[J].International Journal of Computer Vision,2006,70(1)∶77-90.
[8]LI Y Y.Curvature-aware manifold learning[J].Pattern Recognition,2018,83∶273-286.
[9]TENENBAUM J B,SILVA V D,Langford J C.A Global Geometric Framework for Nonlinear Dimensionality Reduction[J].Science,2000,290∶2319-2323.
[10]邹朋成,王建东,杨国庆.辅助信息自动生成的时间序列距离度量学习[J].软件学报,2013,24(11)∶2642-2655.
[11]Márcio Macedo,Antonio Apolinário.Improved antialiasing for Euclidean distance transform shadow mapping[J].Computers&Graphics,2018,71∶166-179.
[12]Tang J R,Cheng H,Zhao Y.Structured dynamic time warping for continuous hand trajectory gesture recognition[J].Pattern Recognition,2018,80∶21-31.
[13]李正欣,张凤鸣,李克武.一种支持DTW距离的多元时间序列索引结构[J].软件学报,2014,25(3)∶560-575.
[14]CHANG Y,HU C,TURK M.Manifold of Facial Expression[C]//IEEE International Workshop on Analysis and Modeling of Faces and Gestures.IEEE,2003∶28-35.