A development of an augmented registration approach towards Scoliosis’s radiographic images stitching method

摘要

In this paper, an augmented registration approach towards Image stitching method is developed by examining the graciousness of feature based method jointly with direct based methods to identify and stitch a pair of overlapped images is presented. The developed method based on this hybrid approach is investigated in term of performance and accuracy. The ability of the examined method to classify overlapping pairs of images is expounded. It is perceived that the proposed method displays superiority performance to the methods described in the literature on the standard databases. The experiments shown promising and potential results for further development to tackle more complicated stitching issues.