Enhanced Extended-Field-of-View Ultrasound for Musculoskeletal Tissues Using Parallel Computing

Extended-Field-of-View (EFOV) ultrasound (US) generates panoramic ultrasound images based on a procedure of image registration applied to a sequence of collected 2D B-scans. In order to improve the imaging accuracy, we propose a novel image registration method combining scale invariant feature transform (SIFT) and random sample consensus (RANSAC) algorithm for the EFOV sonography. Meanwhile, due to the large amount of image data to be processed, implementation of the sequential algorithm based on the SIFT+ RANSAC is time-consuming. To this end, the technology of distributed data processing is employed and an efficient parallel computation method is designed for imaging human musculoskeletal tissues. The experimental results show that the proposed algorithm for EFOV US can produce panoramic images with improved accuracy in comparison with conventional method, especially when the rotations among the captured images are relatively large. In addition, the proposed parallel method significantly speeds up the generation process of panorama in comparison with the serial method, indicating good merit for clinical applications.