【24h】

Analysis of visual speech signatures

机译:视觉语音签名分析

获取原文

摘要

The visual components of speech are the subject of research in resent years. The motivation of this paper is to suggest the concept of signature for visualized speech. These visual signatures are defined for basic components of mouth movement corresponding to sound units. Therefore, an accurate and unified database for perception and generation of visual speech could be organized. Toward this aim, mathematical tools and processes are suggested for concluding the speech related visual units. The lip geometry is parameterized with feature points. The main interpolation tool has been chosen as the Barycentric Lagrange interpolation (BLI). It can include the mouth geometry's structure dynamically during articulation in mathematical form. The outcomes (experimental models) of proposed approach associate mathematical expressions. Such method could be successor of lip tracking and identification algorithm in both lip reading and visual speech animation systems.
机译:语音的视觉组分是遗产岁月的研究主题。本文的动机是建议可视化演讲的签名概念。这些视觉签名定义为对应于声音单元的口径的基本组件。因此,可以组织精确和统一的数据库,用于感知和产生视觉语音。针对此目的,建议对语音相关的视觉单元结束的数学工具和流程。唇部几何与特征点的参数化。主要插值工具已被选为重心拉格朗日插值(BLI)。它可以在数学形式中动态地包括嘴巴几何结构。建议方法的结果(实验模型)助理数学表达。这种方法可以是唇读和视觉语音动画系统中的唇迹跟踪和识别算法的继承者。

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