首页> 外文期刊>Pattern recognition and image analysis: advances in mathematical theory and applications in the USSR >Inverse Problem for the Vocal Tract: Identification of Control Forces from Articulatory Movements
【24h】

Inverse Problem for the Vocal Tract: Identification of Control Forces from Articulatory Movements

机译:声带的反问题:从发音运动中识别控制力

获取原文
获取原文并翻译 | 示例
       

摘要

Various classes of signals controlling articulators in the mathematical model of speech production are investigated, namely: stepwise, continuous and discontinuous, piecewise-linear, piecewise-square, and spline signals. Input data for the inverse problem with respect to the control problem are trajectories of articulatory movements measured with the help of a X-ray microbeam system and cineradiography. The piecewise-linear discontinuous control yields the best approximation of articulatory movements, and the simulated efforts of the correspondent muscles prove to be physiologically attainable. The error of approximation is comparable with the errors in the trajectory measurements. The obtained functional class of controls and the additional con-straints on the values of the control parameters provide a stable and locally unambiguous solution to the dynamic inverse problem for the vocal tract. This justifies the inclusion of our model in a general speech production model.
机译:研究了语音产生数学模型中各种控制发音器的信号,包括:逐步,连续和不连续,分段线性,分段平方和样条信号。与控制问题有关的反问题的输入数据是借助X射线微束系统和X线摄影术测量的关节运动轨迹。分段线性不连续控制产生了最佳的关节运动近似,并且相应肌肉的模拟作用被证明在生理上是可以实现的。近似误差与轨迹测量中的误差相当。所获得的控制功能类别以及对控制参数值的附加约束为声道的动态反问题提供了稳定且局部明确的解决方案。这证明了将我们的模型包含在通用语音生成模型中是合理的。

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