基于VR的言语失用症康复训练系统  

Rehabilitation Training System of VR-based Speech Apraxia

作　　者：焦欢欢 潘志庚 徐舒畅[1] 袁庆曙[1] Jiao Huanhuan;Pan Zhigeng;Xu Shuchang;Yuan Qingshu(School of Information Science and Engineering,Hangzhou Normal University,Hangzhou 311121,China;Alibaba Business School,Hangzhou Normal University,Hangzhou 311121,China;Virtual Reality and Intelligent Systems Research,Hangzhou Normal University,Hangzhou 311121,China;Guangzhou nine Digital Technology Co.,Ltd.,Guangzhou 510000,China)

机构地区：[1]杭州师范大学信息科学与工程学院,浙江杭州311121 [2]杭州师范大学阿里巴巴商学院,浙江杭州311121 [3]杭州师范大学虚拟现实与智能系统研究院,浙江杭州311121 [4]广州玖的数码科技有限公司,广东广州510000

出　　处：《系统仿真学报》2019年第11期2344-2349,共6页Journal of System Simulation

基　　金：国家重点研发计划(2017YFB1002803);广州创新创业领军团队项目(CXLJTD-201609);浙江省健康智慧厨房系统集成重点实验室开放课题(2017F04)

摘　　要：针对传统言语失用症康复训练中患者主动训练积极性逐渐减弱的问题,设计了一种基于虚拟现实的言语失用症康复训练系统。采用虚拟人嘴型动画模拟汉语中26个音节发音嘴型,并用于学习模块。利用孤立字语音识别算法识别26个音节,提高识别的准确率。系统还用训练模块、场景模块辅助患者康复训练。结果表明:系统中的虚拟人发音过程能够逼真地模拟真实发音器官的运动,对提高患者的语言能力具有积极作用。Aiming at the problem that the enthusiasm of active training in patients with traditional speech apraxia rehabilitation is gradually weakened, a virtual reality-based rehabilitation training system for speech apraxia is designed. The system simulates the 26 monosyllabic pronunciation mouths shapes made by virtual human in Chinese and uses it for the learning module. It is proposed to identify 26 monosyllabic words by using the isolated word speech recognition algorithm, which improves the accuracy of speech recognition. The system includes a training module and a scene module for assisting the patient’s pronunciation rehabilitation training. The evaluation results show that the virtual human articulation process in the system can realistically simulate the movement of real articulation organs, which plays a positive role in improving the patients’ language ability.

关 键 词：虚拟现实 言语失用症 虚拟人嘴型动画 孤立字语音识别 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]