可视语音合成技术的3D通信关键技术研究  

Research on Key Technologies of 3D Communication Using Visual Speech Synthesis

作　　者：王文敏 李胜辉 Wang Wenmin;Li Shenghui(China Mobile Online Service Co.,Ltd.,Zhengzhou 450000,China)

机构地区：[1]中移在线服务有限公司,郑州450000

出　　处：《单片机与嵌入式系统应用》2021年第11期42-46,共5页Microcontrollers & Embedded Systems

摘　　要：针对3D通信中人物模型面部表情难以控制、音频和动作数据同步性较差的问题,本研究基于可视语音合成技术的3D通信平台,用户终端对接收到的数据进行人物化表达,通过虚拟人物模型的肢体动作和面部表情传递情感和意图。基于OpenGL工具开发,将纹理展开图作为映射材质,提高了虚拟人物模型的真实性。实验结果显示,本研究输出的语音数据正确率最高可达到99%,面部模型的动作数据与真实运动数据最为接近。It is difficult to control the facial expression of the character model in 3D communication,and the synchronization of audio and motion data is poor.This study is based on the 3D communication platform of visual speech synthesis technology,and the user terminal is able to personalize the data received and express the emotion and intention in the information transmitted through the body movements and facial expressions of the virtual character model.Based on OpenGL tool development,texture expansion map is used as the mapping material to improve the authenticity of the virtual character model.The experiment results show that the accuracy of the speech data output in this study is up to 99%,and the motion data of the face model is closest to the real motion data.

关 键 词：3D通信 可视语音合成 网格抽象化 音频切割 OPENGL 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]