Determination of three-dimensional structured objects, vascular structures, and imaging geometry from single-plane and biplane projection images
Nazareth, Daryl P
MetadataShow full item record
Three-dimensional (3D) vessel trees can provide useful visual and quantitative information during interventional procedures. To calculate the 3D vasculature and improve these measurements, we have developed methods for the determination of geometric parameters from single-plane and biplane projection images. Our single-plane technique provides an accurate estimation of the magnification and orientation of objects of known dimensions in vessels by comparing measurements in the images with those in simulated images of modeled objects. Our biplane technique calculates the transformation relating the imaging systems (i.e., the rotation matrix R and the translation vector t ) and requires only the identification of approximately corresponding vessel regions in the two images. Initial estimates of R and t are refined using an optimization method. The objective function to be minimized is based on the amount of overlap of corresponding vessel regions in the two images. The 3D vasculature is then obtained from the optimal R and t using triangulation. The accuracy of the 3D vasculature calculations may be further improved when a calibration object, such as a stent, is present in the vasculature and the biplane images, if the required user-indicated points in the stent are highly accurate. We have modified the above biplane technique to incorporate information provided by the stent, by including three additional terms in the objective function. These techniques were evaluated using simulated and phantom images. The single-plane technique provided accuracies of 1% in magnification and 2 degrees in orientation. The biplane technique provided accuracies of 1% and 1 degree, respectively, which was reduced to 0.3% and 0.5 degrees in simulations when a calibration object was present. The results of the biplane technique applied to the phantom indicated that inaccuracies in user indication of the calibration object may propagate into the errors in the 3D vessel tree reconstruction. Thus, overdependence on the calibration object may degrade accuracy. As a result of these techniques, more accurate values of diameters and lengths will be available during surgical procedures. In addition accurate 3D measures may be useful in clinical decision making, such as in assessing vessel tortuousity and access, during interventional procedures.