Dynamic Real-Time 3-Dimensional Model of the Tongue’s Motion

Patients undergoing traditional speech therapy often have difficulty grasping the therapist’s audible cues for the correct intended movement of the tongue. This lack of clarity is attributed to the absence of visual biofeedback. Therefore, a model to visualize the tongue’s real-time motion could assist in providing the necessary biofeedback required in speeding up the process of the therapy. Hence, we have developed a dynamic 3-Dimensional (3D) model of the human tongue using the Unity software platform (Unity Technologies, San Francisco). This model is a standalone platform that could also be used in conjunction with the discreet oral wearable device developed by our team. The model was developed from static Magnetic Resonance (MR) images of the tongue, collected at UIC Center for Magnetic Resonance Research. 34 series of MR images of the static position of the tongue during various static tasks were captured. In order to create the 3D model, first the MR images were segmented. Then, a mesh of the tongue was developed and optimized for all the shapes that the tongue could assume. The meshes were then exported to Unity and were animated. The model of the motion of the tongue was enabled by moving a set of splines attached to the mesh. Furthermore, a User Interface (UI) was developed to interact with the model. The resulted model allows for reproduction of movements associated with 24 English phonemes and the words composed of these 24 phonemes. The model could interact in real-time with the oral wearable device in order to create live models of the motion. The model is also capable of reproducing the tongue’s motion using offline data from the wearable device.