Department：School of Artificial Intelligence
Main courses：Speech Signal Processing
Tutor type：doctoral supervisor
Research field：Speaking is a unique capacity of humans, which involves physiological and acoustic mechanisms of a human body across many stages of linguistic, neural sensorimotor, peripheral bio mechanic, and aero-acoustic systems. Studies of speech production aim at exploring those mechanisms by interdisciplinary approach combining phonetics, physiology, engineering, and medicine.
Current Research Topics
Visualization of Speech Production with Magnetic Resonance Imaging (MRI)
MRI is the most powerful technique to observe the speech organs and the vocal tract. Various imaging techniques are available to use, including 2D or 3D static images, partial-3D motion images, and 2D real-time images during production of stationary speech gestures or running speech. The vocal tract (upper airway) can be extracted from MRI data to simulate vowel production by means of a linear electrical analog method or 3D acoustic simulation techniques.
Modeling Speech Production Processes
Basic mechanisms of speech production often need to be clarified by model-based simulations. In physiological processes, models of speech organs such as the tongue and lips are useful to understand the mechanisms, and in acoustic processes, models of the acoustic phenomena in the vocal tract help us understand the physical mechanisms. The neural processes also need modeling because of the complexity as a whole.
Developing Instrumentation Techniques for Speech Production Studies
Technical difficulty in instrumentation has been the wall for examining human speech production. Multi-sensory instrumentation techniques are requested for this purpose using non-invasive experiments. Development of several new instruments has been conducted to build a new system for measuring speech production.
Research Interests：Speech Production