Automated reconstruction of standing posture panoramas from multi-sector long limb x-ray images
暂无分享,去创建一个
Due to the digital X-ray imaging system's limited field of view, several individual sector images are required to capture the posture of an individual in standing position. These images are then “stitched together” to reconstruct the standing posture. We have created an image processing application that automates the stitching, therefore minimizing user input, optimizing workflow, and reducing human error. The application begins with pre-processing the input images by removing artifacts, filtering out isolated noisy regions, and amplifying a seamless bone edge. The resulting binary images are then registered together using a rigid-body intensity based registration algorithm. The identified registration transformations are then used to map the original sector images into the panorama image. Our method focuses primarily on the use of the anatomical content of the images to generate the panoramas as opposed to using external markers employed to aid with the alignment process. Currently, results show robust edge detection prior to registration and we have tested our approach by comparing the resulting automatically-stitched panoramas to the manually stitched panoramas in terms of registration parameters, target registration error of homologous markers, and the homogeneity of the digitally subtracted automatically- and manually-stitched images using 26 patient datasets.
[1] Thorsten M. Buzug,et al. An Algorithm for Automatic Stitching of CR X-ray Images , 2007 .
[2] André Gooßen,et al. A Stitching Algorithm for Automatic Registration of Digital Radiographs , 2008, ICIAR.
[3] David H. Foos,et al. Fully automatic and reference-marker-free image stitching method for full-spine and full-leg imaging with computed radiography , 2004, SPIE Medical Imaging.
[4] Leo Joskowicz,et al. Long bone panoramas from fluoroscopic X-ray images , 2004 .