Chinese Mandarin Speech Learning Software Based On 3D Motion Capture

In this thesis,the three-dimensional motion capture system is used to collect the speech data of Chinese Mandarin and the coordinate data of face feature points.The pronunciation characteristics of Mandarin Chinese are studied and analyzed,and the design of Chinese Mandarin speech learning software is designed.Thereby achieving the purpose of voice teaching.The thesis of the research work mainly includes the following three parts:(1)Using the Vicon Cara and Dynamixyz facial capture systems to retrofit them to obtain a motion capture system that can simultaneously capture audio and video.Firstly,face facial feature point data and voice data are collected.Secondly,the face feature points are processed by the device's software CaraPost.Finally,the processed three-dimensional face coordinate points are got using Motion Builder.(2)Construct a face model using Matlab development tools.Firstly,the face mesh model method is used to construct the 3D face feature points.Secondly,because the constructed face lip curve is not vivid and natural,the cubic spline interpolation method is used to smooth the inner and outer lip contours to achieve smoothness;finally,the synthesis of 3D face and voice data,using the Premiere tool to render the face and voice,and use the cut tool to segment the syllables,and finally get all the dynamic 3D face pronunciation video library.(3)Realize the design of Mandarin Chinese speech learning software by using Director multimedia development tools.Firstly,according to the design,collecting and producing needed materials,and the software's three-level interface is designed.Secondly,the functional part of the software is assembled and built,and the Lingo program language is used to control the change of the driver interface and the realization of the function;Finally,test the developed software to check the stability and practicability of the software.