Innovative Voice Analysis System To Be Developed for Language and Speech Therapy
March 20, 2024-GTU Office of Press and Public Relations
A practical software system is to be developed and made ready for widespread use in collaboration with GTU and UHS to assist in the diagnosis and therapy of language and speech disorders.
Support from TÜSEB
Researchers from Gebze Technical University (GTU) and University of Health Sciences (HBU) have become entitled to receive support for the "USradioAI" project in the field of language and speech therapy under TÜSEB's (Health Institutes of Türkiye) Artificial Intelligence-Based Medical Image Processing Project Call in Priority R&D Areas (2023-C1-YZ IMAGE).
The project titled "USradioAI: Millimeter Wave (mmWave) Based Voice Analysis System Trained with Ultrasonic Tongue Images for Language and Speech Therapy" includes Prof. Dr. Yusuf Sinan Akgül and Research Assistant İlhan Aytutuldu from the Department of Computer Engineering at GTU and Medical Doctor Tunahan Refik Dumlu from the Internal Medicine Department of Istanbul Kartal Dr. Lütfi Kırdar Health Application and Research Center, University of Health Sciences (UHS).
Within the scope of the project, the aim is to develop a mmWave radar-based voice analysis system trained with ultrasonic tongue images. It is aimed for this system to be a practical software system that will assist in the diagnosis and therapy of language and speech disorders. The developed system has been designed to be less costly and more practical than ultrasound devices commonly used in clinics, and it is expected to operate with high accuracy.
Project Coordinator Prof. Dr. Yusuf Sinan Akgül and his team designed a system that will incorporate vocal tract vibration, lip movements, and facial information into the analysis process using mmWave radar data. Thus, real-time monitoring of patients' facial, lip, and vocal tract vibrations will be possible, providing feedback for the correct pronunciation of words or proper use of muscles.
Furthermore, visualization will be provided in the International Phonetic Alphabet chart representing the placement of sound in the oral cavity using the system developed in the project, aiming to make the therapy process more effective and personalized.
The project aims to develop a system that takes into account the spatial and temporal dimensions of language positions, facial, lip, and vocal tract vibrations, thus aiming to develop a system that provides more consistent and accurate phonetic detection over time. Thus, it is aimed to increase the accuracy rates in the sound analysis process and achieve more effective results.
This project, carried out in collaboration with Gebze Technical University and the University of Health Sciences, aims to make significant progress in the field of language and speech therapy.
