Gradient-based 2-D/3-D rigid registration of fluoroscopic X-ray to CT

Abstract

Abstract-We present a gradient-based method for rigid registration of a patient preoperative computed tomography (CT) to its intraoperative situation with a few fluoroscopic X-ray images obtained with a tracked C-arm. The method is noninvasive, anatomybased, requires simple user interaction, and includes validation. It is generic and easily customizable for a variety of routine clinical uses in orthopaedic surgery. Gradient-based registration consists of three steps: 1) initial pose estimation; 2) coarse geometry-based registration on bone contours, and; 3) fine gradient projection registration (GPR) on edge pixels. It optimizes speed, accuracy, and robustness. Its novelty resides in using volume gradients to eliminate outliers and foreign objects in the fluoroscopic X-ray images, in speeding up computation, and in achieving higher accuracy. It overcomes the drawbacks of intensity-based methods, which are slow and have a limited convergence range, and of geometry-based methods, which depend on the image segmentation quality. Our simulated, in vitro, and cadaver experiments on a human pelvis CT, dry vertebra, dry femur, fresh lamb hip, and human pelvis under realistic conditions show a mean 0.5-1.7 mm (0.5-2.6 mm maximum) target registration accuracy.Index Terms-Fluoroscopic X-ray to CT registration, gradient based, image registration, 2D/3D rigid registration.

A bounded iterative closest point method for minimally invasive registration of the femur

“…Fluoroscopic methods can either be used to track fiducial markers (20) or bone anatomy directly (21), and typically use a fixed position or tracked C-arm fluoroscope. While images can be much cleaner than those produced by ultrasound imaging, the use of X-rays does present the drawback of increasing the patient's radiation exposure.…”

Vision Based Robotic Navigation: Application to Orthopedic Surgery

“…These methodologies can be grouped as feature based registration techniques (Besl and McKay. 1992;Taylor, Mittelstadt et al 1994;Lavallee, Szeliski et al 1995;Livyatan, Yaniv et al 2003;Tomaževic, Likar et al 2003) or as intensity based registration techniques (Penney, Batchelor et al 2001;Zollei, Grimson et al 2001;Rohlfing and Maurer 2002). Feature based approaches rely on the presence and identification of natural landmarks in the input datasets in order to determine the best alignment.…”

“…Surface based (Besl and McKay. 1992;Lavallee, Szeliski et al 1995;Livyatan, Yaniv et al 2003;Tomaževic, Likar et al 2003) feature registration approaches utilize the outer surfaces of an object of interest (extracted from a 3D model), and contours that outline the object on one or more 2D images. Registration is concerned with finding the object pose that minimizes the distance between surface and the contour.…”

“…Another methodology is to project 3D surface points onto the 2D plane and then attempt to minimize the distance between the contour of the projected image and the initial 2D contour. The work done by (Lavallee, Szeliski et al 1995;Livyatan, Yaniv et al 2003;Tomaževic, Likar et al 2003) has presented another methodology where projections of the volumetric data gradients are computed and compared with X-ray image gradients. The volumetric data pose is adjusted to minimize this gradient difference.…”

Surface/Volume-Based Articulated 3D Spine Inference through Markov Random Fields

“…For example in [4,5], 3D models obtained from CT or MR were registered to 2D X-ray and fluoroscopic images using gradient amplitudes for optimizing the correspondence of single bone structures. Similar objective functions using surface normals from statistical PDMs [6] were applied for the femur.…”