Based upon motion capture,a semi-automatic technique for fast facial animation was implemented. While capturing the facial expressions from a performer,a camera was used to record her /his front face as a texture map....Based upon motion capture,a semi-automatic technique for fast facial animation was implemented. While capturing the facial expressions from a performer,a camera was used to record her /his front face as a texture map. The radial basis function( RBF) technique was utilized to deform a generic facial model and the texture was remapped to generate a personalized face.Partitioning the personalized face into three regions and using the captured facial expression data,the RBF and Laplacian operator,and mean-value coordinates were implemented to deform each region respectively. With shape blending,the three regions were combined together to construct the final face model. Our results show that the technique is efficient in generating realistic facial animation.展开更多
提出了一种在粒子滤波框架下的结合在线外观模型(online appearance model,OAM)和柱状人头模型(cylinder head model,CHM)的人脸三维运动跟踪方案,具体包括:1)融合多种观测信息来降低OAM的光照敏感性和个体相关性;2)针对OAM适合跟踪局...提出了一种在粒子滤波框架下的结合在线外观模型(online appearance model,OAM)和柱状人头模型(cylinder head model,CHM)的人脸三维运动跟踪方案,具体包括:1)融合多种观测信息来降低OAM的光照敏感性和个体相关性;2)针对OAM适合跟踪局部运动但在大姿态下会跟踪失败的问题,将OAM与适合于大姿态下全局运动跟踪的CHM结合起来,在当前帧将CHM匹配得到的全局运动参数作为OAM匹配的初始值,将OAM匹配得到的人脸运动参数作为下一帧CHM匹配的初始值;3)基于局部优化和改进重采样来改进粒子运动滤波策略.实验表明:该系统在大姿态、表情剧烈变化、遮挡和强光照下能得到较好的跟踪效果,且OAM+CHM的跟踪正确率高于OAM的24%,OAM+CHM的姿态跟踪范围大于OAM的11%.主观实验表明:由跟踪得到的人脸运动参数合成的虚拟人脸具有较高的辨识度.展开更多
目的应用面部三维动态定量分析系统(3-Dimensional Analysis System for Facial Motion,3D-ASFM)对面神经部分切除术后的患者进行测量,探讨修复与未修复患者面部静态和动态指标的变化。方法14例面神经部分切除患者,分为神经修复组(9例)...目的应用面部三维动态定量分析系统(3-Dimensional Analysis System for Facial Motion,3D-ASFM)对面神经部分切除术后的患者进行测量,探讨修复与未修复患者面部静态和动态指标的变化。方法14例面神经部分切除患者,分为神经修复组(9例)和未修复组(5例)。对所有患者进行半年以上跟踪随访,中位随访期13个月。所有患者术前及术后均进行面部运动主观评估(House-Brackmann分级和Sunnybrook评分)和面部三维动态测量。比较三种评估结果相关性。比较修复组与未修复组患者术前、术后各评估结果,对修复组患者各面部区域进行分析。结果3D-ASFM分析结果分别与House-Brackmann分级(H-BGS)、Sunnybrook评分(SFGS)进行秩相关分析,相关系数分别为-0.707(P<0.05)和0.798(P<0.05)。面瘫越严重,3D-ASFM与主观评价系统的相关性越差。术后H-BGS、SFGS总分均存在统计学差异(P<0.05),修复组好于未修复组,3D-ASFM总分在两组间无统计学差异(P>0.05),但鼻旁运动好于未修复者。修复组患者术前、术后H-BGS、SFGS、3D-ASFM总分均无统计学差异(P>0.05)。所有患者健侧动态指标中,上面部运动较术前减小,下面部运动较术前增加。结论面部三维动态定量分析结果与传统主观评价系统有较好的相关性。神经修复对面中部运动影响最大。面神经切除手术不仅造成同侧面部运动改变,也会影响对侧面部运动。展开更多
基金Youth Foundation of Higher Education Scientific Research of Hebei Province,China(No.2010228)Foundation for Returned Overseas Scholars of Hebei Province,China(No.C2013003015)
文摘Based upon motion capture,a semi-automatic technique for fast facial animation was implemented. While capturing the facial expressions from a performer,a camera was used to record her /his front face as a texture map. The radial basis function( RBF) technique was utilized to deform a generic facial model and the texture was remapped to generate a personalized face.Partitioning the personalized face into three regions and using the captured facial expression data,the RBF and Laplacian operator,and mean-value coordinates were implemented to deform each region respectively. With shape blending,the three regions were combined together to construct the final face model. Our results show that the technique is efficient in generating realistic facial animation.
文摘提出了一种在粒子滤波框架下的结合在线外观模型(online appearance model,OAM)和柱状人头模型(cylinder head model,CHM)的人脸三维运动跟踪方案,具体包括:1)融合多种观测信息来降低OAM的光照敏感性和个体相关性;2)针对OAM适合跟踪局部运动但在大姿态下会跟踪失败的问题,将OAM与适合于大姿态下全局运动跟踪的CHM结合起来,在当前帧将CHM匹配得到的全局运动参数作为OAM匹配的初始值,将OAM匹配得到的人脸运动参数作为下一帧CHM匹配的初始值;3)基于局部优化和改进重采样来改进粒子运动滤波策略.实验表明:该系统在大姿态、表情剧烈变化、遮挡和强光照下能得到较好的跟踪效果,且OAM+CHM的跟踪正确率高于OAM的24%,OAM+CHM的姿态跟踪范围大于OAM的11%.主观实验表明:由跟踪得到的人脸运动参数合成的虚拟人脸具有较高的辨识度.
文摘目的应用面部三维动态定量分析系统(3-Dimensional Analysis System for Facial Motion,3D-ASFM)对面神经部分切除术后的患者进行测量,探讨修复与未修复患者面部静态和动态指标的变化。方法14例面神经部分切除患者,分为神经修复组(9例)和未修复组(5例)。对所有患者进行半年以上跟踪随访,中位随访期13个月。所有患者术前及术后均进行面部运动主观评估(House-Brackmann分级和Sunnybrook评分)和面部三维动态测量。比较三种评估结果相关性。比较修复组与未修复组患者术前、术后各评估结果,对修复组患者各面部区域进行分析。结果3D-ASFM分析结果分别与House-Brackmann分级(H-BGS)、Sunnybrook评分(SFGS)进行秩相关分析,相关系数分别为-0.707(P<0.05)和0.798(P<0.05)。面瘫越严重,3D-ASFM与主观评价系统的相关性越差。术后H-BGS、SFGS总分均存在统计学差异(P<0.05),修复组好于未修复组,3D-ASFM总分在两组间无统计学差异(P>0.05),但鼻旁运动好于未修复者。修复组患者术前、术后H-BGS、SFGS、3D-ASFM总分均无统计学差异(P>0.05)。所有患者健侧动态指标中,上面部运动较术前减小,下面部运动较术前增加。结论面部三维动态定量分析结果与传统主观评价系统有较好的相关性。神经修复对面中部运动影响最大。面神经切除手术不仅造成同侧面部运动改变,也会影响对侧面部运动。