Comparison of Object Tracking Algorithms for Larynx Phantom Movements in Ultrasound Videos
Authors: Christian Kleiner, Peter Birkholz
Abstract:
A larynx phantom based on gelatine and an excised pig larynx was built. Larynx phantom movement was determined in RGB video for reference. Additionally, larynx phantom movement was estimated in ultrasound video using two different tracking algorithms based on optical flow and template matching, respectively. In ultrasound video, high tracking accuracies of up to 97% were found for both tracking algorithms. This supports the validity of object tracking in laryngeal ultrasound video as a method to measure larynx height.