An Alternative Technique for Imaging Registration in IVUS Images

Image registration technique has played a major role due to the use of data from different medical imaging modalities in which your main goal is to find anatomical and/or functional matches in two or more images. SIFT algorithm was used to extract a set of descriptors in IVUS images of two different pullbacks and to determine the number of common features between each pair of images. Similarity map was created to establish correspondences between two PCI images. Our method, with saltatory condition, was proved be a good choice in the experiments and in the real case, showing great stability in the establishment of correspondences and reducing overall time used for the diagnosis and/or monitor the development of coronary heart disease.