CT from an Unmodified Standard Fluoroscopy Machine Using a Non-reproducible Path

Abstract

3D reconstruction from image data is required in many medical procedures. Recently, the use of fluoroscopy data to generate these 3D models has been explored. Most existing methods require knowledge of the scanning path either from precise hardware, or pre-calibration procedures. We propose an alternative of obtaining this needed pose information without the need of additional hardware or pre-calibration so that many existing fluoroscopes can be used.

Our method generates 3D data from fluoroscopy collected along a non-repeatable path using cone-beam tomographic reconstruction techniques. The novelty of our approach is its application to imagery from existing fluoroscopic systems that are not instrumented to generate pose information or collect data along specific paths. Our method does not require additional hardware to obtain the pose, but instead gathers the needed object to camera pose information for each frame using 2D to 3D model matching techniques [1-3]. Metallic markers are attached to the object being imaged to provide features for pose determination. Given the pose, we apply Grangeat’s cone-beam reconstruction algorithm to recover the 3D data.

In developing this approach, several problems arose that have not been addressed previously in the literature. First, because the Radon space sampling is different for each scan, we cannot to take advantage of a known Radon space discretization. Therefore we have developed a matching score that will give the best Radon plane match for the resampling step in Grangeat’s approach [4]. Second, although we assume Tuy’s condition [5] is satisfied, there are sometimes data gaps due to discretization. We have developed a method to correct for these gaps in the Radon data.