Parallelism of Extended-Field-of-View Sonography based on Scale Invariant Feature Transform

EFOV sonography can acquire a full field of vision, which can help doctors to make more objective and accurate diagnosis. This paper explores parallel implementation techniques for Extended-Field-of-View (EFOV) Sonography based on Scale Invariant Feature Transform (SIFT). In our method, the SIFT algorithm is used to more accurately register the images. Bidirectional matching algorithm is chosen to improve the accuracy of image registration. For real-time applications, we propose a parallel computing approach to speed up the generation of panorama. Based on the cluster of PCs, we erect a parallel computing platform and perform preliminary experiments. The results demonstrate the effectiveness of our approach by comparing to traditional methods.