Kernel-based Tracking in Ultrasound Sequences of Liver

Objective - Object tracking in 2D ultrasound sequences of liver to infer real-time respiratory organ movement and offer motion compensation in image-guided abdominal interventions. Methods - A kernel-based tracking algorithm that is adaptive to scale and orientation changes of the tracking target is applied to 54 vessel targets in 21 ultrasound sequences acquired from volunteers under free breathing. Tracking performance is evaluated based on manually annotated ground truth information. Results - Tracking results show that the algorithm is able to track the assessed targets in a precise and robust manner in real-time performance. The overall mean tracking error is 1.43 (STD 1.22 mm).     «

Objective - Object tracking in 2D ultrasound sequences of liver to infer real-time respiratory organ movement and offer motion compensation in image-guided abdominal interventions. Methods - A kernel-based tracking algorithm that is adaptive to scale and orientation changes of the tracking target is applied to 54 vessel targets in 21 ultrasound sequences acquired from volunteers under free breathing. Tracking performance is evaluated based on manually annotated ground truth information. Result...     »