自适应人脸语音动画技术-计算机应用研究.PDFVIP

1·48 · 计算机应用研究 2006 年 * 自适应人脸语音动画技术 徐 莹1, 3 , 何本阳2 ( 1. 江苏大学 计算机系, 江苏 镇江 212013; 2. 同济大学 计算机系, 上海 200092; 3. 镇江高等专科学校 电子 信息系, 江苏 镇江 212003) 摘 要: 介绍了一个基于特征多边形网格模型和肌肉模型的人脸语音动画系统。首先从 3D MAX 导出的 ASE 文件建立人脸多边形网格模型, 在四边形网格中定义了眼睛、鼻子、眉毛、额头、面颊、上下颚等特征点。基于特 征点把人脸网格模型构造为线性弹性模型, 并进行纹理贴图, 最后在建立的肌肉模型基础上实现人脸语音动画。 关键词: 特征点 ; 线性弹性模型; 纹理映射; 肌肉模型; 语音同步动画 中图法分类号: TP391 文献标识码 : A 文章编号 : 1001- 3695( 2006) 03- 0148- 03 Adaptive Facial Animation of Speech Synchronization 1, 3 2 XU Ying , HE Ben-yang ( 1. Dept. of Computer Science Engineering, Jiangsu University, Zhenjiang Jiangsu 212013 , China; 2. Dept. of Computer Science Engi- neering, Tongji University, Shanghai 200092, China; 3. Dept. of Electricity Information, Zhenjiang College, Zhenjiang Jiangsu 212003 , China) Abstract: A virtual speech synchronization system for facial animation based on linear spring mesh model muscle model is de- signed. By importing ASE files to construct a polygonal mesh model, on which many feature points are added in eyes, nose, eyebrow, forehead, jowl, maxilla and chin. Then, this feature-based polygonal mesh model is reconstructed into a linear spring mesh model. After texture mapping and adding muscle model, highly realistic facial expression and speech-synchronized facial animation are achieved. Key words: Feature Point; Linear Spring Mesh Model; Texture Mapping; Muscle Model; Speech-synchronizing Animation 下颚等特征点。基于特征点把人脸网格模型构造为线性弹性 1 引言