Rehabilitation Training System Of VR-based Speech Apraxia

Apraxia of speech is a language dyskinesia caused by brain damage,often accompanied by the emergence of aphasia.The existence of speech apraxia seriously affects patients' ability of daily living and their ability to return to society.Traditional rehabilitation therapy is usually performed by a speech rehabilitation practitioner in a hospital to demonstrate articulatory movement of the vocal organs and ask patients to imitate.The process is repetitive,monotonous and expensive,which makes it difficult to arouse patients' interest.In recent years,due to its unique advantages: strong immersion,strong interaction,and strong interest,virtual reality technology has been increasingly used in the field of rehabilitation.In this paper,a virtual reality-based rehabilitation training system for speech apraxia is designed to solve the problem that patients' active training initiative gradually decreases in the traditional rehabilitation training for speech apraxia.The system uses virtual human's mouth type animation to simulate the 26 monosyllabic pronunciators' s mouth type in Chinese,and uses it for the learning module;In addition,the training module and the scene module are used to assist the patients in pronunciation rehabilitation training.This paper first introduces the related technologies of system development tool Unity3 D and interactive devices HTC Vive and Fingo,and then designs a feasible solution by analyzing the interactive requirements of the system.After that,this paper focuses on the four modules of the system: Teaching module,learning module,training module,and scene module for functional design and interaction design.Specific research content includes:1.Virtual human mouth type drive,using DFFD(Dirichlet free-free deformation),by controlling the displacement of the main feature points to obtain the deformed virtual human monosyllabic mouth shape,simulating human skin movement,thus 3D virtual person can produce the pronunciation movements that conforms to the physiological characteristics of the human body.2.Speech recognition technology,using the isolated word speech recognition algorithm based on dynamic time warping(DTW)to identify the monosyllabicsyllables,the speech recognition of iFLYTEK open platform is used to recognize words and sentences.3.Design and implement a virtual rehabilitation training system.Through the learning module,training module and scene module,the system allows the patient to carry out rehabilitation training on the pronunciation of the rhymes,the word learning and the application of the learned words in the corresponding scenes.The experimental results show that the virtual rehabilitation training system studied in this paper is effective for the treatment of patients' organes,vocal training and speech mode training,and has good application value.