...
首页> 外文期刊>電子情報通信学会技術研究報告. パターン認識·メディア理解. Pattern Recognition and Media Understanding >Construction of 3D facial feature models considering facial muscle movements using two photographs
【24h】

Construction of 3D facial feature models considering facial muscle movements using two photographs

机译:使用两张照片构建考虑面部肌肉运动的3D面部特征模型

获取原文
获取原文并翻译 | 示例
           

摘要

A method for constructing facial images from arbitrary directions and expressions for use in multimedia communication applications is proposed. Both the geometric deformations and physical movements of the facial muscles are considered in this method, and enables the construction of realistic facial images in a simple way. Two orthogonal images of an expressionless face are acquired, and feature points are selected to construct a three dimensional polygon model to describe the shape of the human face. Elastic springs are placed on the model to express the deformations of facial components, and polygon vertices near the springs' dynamic points are displaced accordingly to construct changes in expressions naturally. Since these spring movements do not depend on a specific facial structure, the model could easily be adapted to any given person's face. The method was implemented on an inexpensive PC, and simulation results showed its effectiveness.
机译:提出了一种从任意方向和表情构造面部图像以用于多媒体通信应用的方法。在这种方法中,考虑了面部肌肉的几何变形和物理运动,并且可以以简单的方式构造逼真的面部图像。获取无表情的面部的两个正交图像,并选择特征点以构建三维多边形模型来描述人脸的形状。将弹性弹簧放置在模型上以表达面部组件的变形,并相应地移动弹簧动态点附近的多边形顶点以自然地构造表情变化。由于这些弹簧运动不依赖于特定的面部结构,因此该模型可以轻松地适应任何给定人的脸部。该方法在便宜的PC机上实现,仿真结果表明了其有效性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号