Abstract
Analyzing the motion of the human tongue surface provides valuable information about speech and swallowing. One method to analyse this motion is to acquire two‐dimensional ultrasound images and extract the tongue surface contours from them. Quantitative and statistical analysis of these extracted contours is made difficult because of the absence of physical fleshpoint markers on them. In this research, this problem is overcome by pre‐processing the contours using Kriging. Pre‐processing includes extrapolating and resampling the contours on a regular spatial grid. The preprocessed contours can then be visualized as spatiotemporal surfaces. A dedicated user interface called SURFACES is designed to aid in the generation, visualization and quantitative comparisons of these spatiotemporal surfaces.
Acknowledgments
This research was supported in part by a grant (R01 01758) from National Institutes of Health of United States of America.