Mapping palatal shape to electromagnetic articulography data: An approach using 3D scanning and sensor matching

Area of Science:

• Speech Science
• Biomedical Engineering
• Articulatory Phonetics

Background:

• Accurate three-dimensional (3D) palatal shape data is crucial for understanding speech articulation.
• Existing methods for integrating palatal shape with articulatory tracking data have limitations.

Purpose of the Study:

• To develop and validate a method for superimposing the palatal shape onto 3D electromagnetic articulography (EMA) data.
• To improve the accuracy of speech articulation analysis by integrating detailed palatal geometry.

Main Methods:

• A custom biteplate with dental impression tray and EMA sensors was utilized to capture palatal shape and sensor positions.
• The biteplate was 3D scanned, and the resulting palatal data was mapped to EMA recordings.
• Mapping involved aligning sensor positions from the 3D scan with the EMA data.

Main Results:

• The proposed method successfully superimposed palatal shape onto 3D EMA data.
• The average distance between the mapped palate and EMA traces was approximately 1 mm across nine speakers.
• This accuracy was found to be comparable to the inherent measurement error of EMA.

Conclusions:

• The developed method provides a reliable way to integrate detailed palatal morphology with dynamic speech movement data.
• This technique offers improved spatial accuracy for articulatory speech research.
• The findings suggest this method can enhance the understanding of speech production mechanisms.