Related Experiment Video
Updated: Jun 6, 2026
05:33
Three-Dimensional Shape Modeling and Analysis of Brain Structures
Published on: November 14, 2019
Using statistical deformable models to reconstruct vocal tract shape from magnetic resonance images.
M J M Vasconcelos1, S M Rua Ventura, D R S Freitas
1Faculty of Engineering, University of Porto, Institute of Mechanical Engineering and Industrial Management, Porto, Portugal.
Summary
This study introduces a novel method using statistical deformable models to reconstruct vocal tract shapes for Portuguese European speech sounds from MRI scans. This technique enhances understanding of speech production dynamics for rehabilitation and simulation.
Area of Science:
- Medical Imaging
- Speech Science
- Computational Anatomy
Background:
- Speech production mechanisms are complex and studied using Magnetic Resonance Imaging (MRI) for vocal tract morphology.
- Statistical deformable models, particularly Point Distribution Models (PDMs), are effective for analyzing anatomical structures in medical images.
Purpose of the Study:
- To evaluate the suitability of PDMs for characterizing and reconstructing the vocal tract shape during Portuguese European (EP) speech sound articulation using MRI.
- To establish a novel approach for analyzing dynamic speech events in vocal tract shape.
Main Methods:
- A PDM was constructed from MR images of sustained articulation of 25 EP speech sounds.
- The PDM's capacity to characterize vocal tract shape deformation during sound production was analyzed.
- The model was applied to reconstruct five EP oral vowels and EP fricative consonants.
Main Results:
- The study successfully built a PDM from MRI data of EP speech sounds.
- The PDM effectively characterized vocal tract shape dynamics during sustained articulations.
- Reconstruction of EP vowels and fricatives using the PDM was achieved.
Conclusions:
- This work represents the first approach to characterize and reconstruct vocal tract shape from MRI using PDMs for EP speech sounds.
- The PDM technique provides enhanced understanding of dynamic speech events in sustained articulations.
- Findings are relevant for speech rehabilitation and simulation applications.
