UATracker: A tool for ultrasound data management

Abstract

This presentation introduces TraceTracker, a tool for efficiently managing language ultrasound data. Ultrasound imaging of the tongue is used for analyzing the articulatory features of speech sounds. Most analyses involve finding data points from individual images. The number of image frames and the volume of secondary data associated with them tend to grow quickly in speech analysis studies of this type, making it very hard to handle them manually. TraceTracker is a data management tool for organizing, modifying, and performing advanced searches over ultrasound tongue images and the data associated with those images. The setup operation of the program automatically iterates through file systems and generates a comprehensive database containing the image files and information such as the speaker, the video each frame is extracted from, an index, how they have been traced, etc. The program also automatically reads Praat format TextGrid files and associates specific image frames with the corresponding words and speech segments based on the annotations in the grids. Once the database is populated, TraceTracker can be used to tag images, generate copies, and perform advanced search operations over the images based on the aforementioned criteria including the specific sequence of segments in which it lies.