| dc.description.abstract | The aim of this study is to develop a 3D finite element mechanical tongue model based on the reconstructed MR images of the human tongue. In this study, we used the second and fifth order Mooney-Rivlin hyperelastic strain-energy function that Gerard had used in his paper as the constitutive equation of our tongue model, and assumed that the tongue is incompressible. Based on the reconstructed MR images of the human tongue, we built the tongue model with ANSYS® finite element software. Our tongue model includes seven principle muscles that are involved in the human speech production. They are the Styloglossus muscles, the Hyoglossus muscles, the Genioglossus muscles (also divided into the anterior, middle and posterior three parts.), the Superior longitudinalis muscles, the Inferior longitudinalis muscles, the Transversus muscles and the Verticalis muscles. Once the tongue model was developed, dynamic simulation proceeded on each muscle of the tongue model with the force direction followed the muscle fiber direction. From the results, we could observe that contraction of each muscle in this tongue model had caused correct corresponding deformation on the tongue model, respectively. Finally, we simulated articulatory movements of vowels [u], [a], [i] and consonants [t], [k] with our tongue model. From the results, we also could see that the articulatory movements of our tongue model had conformed to the features of these phonetic articulations. These preliminary results show that our 3D finite element mechanical tongue model could be used as a basis for more phonetic articulatory simulation in the future. | en_US |