Diploma thesis

Abstract

Ultrasound is widely used as a diagnostic and interventional imaging technique in various medical disciplines. Nowadays the latest generation of ultrasound transducers is capable of acquiring three-dimensional images. However, most ultrasound systems still acquire only two-dimensional images. Hence, a favored way to generate three-dimensional ultrasound images is to record a sequence of two-dimensional ultrasound images and the position and orientation of a sensor fixed to the ultrasound probe.

In order to compound the sequence of two-dimensional ultrasound images into a three-dimensional image, one has to determine the rigid transformation from the ultrasound image coordinate system to the coordinate system of the tracking sensor. This process is referred to as spatial calibration.

The next generation of Siemens AcuNav ultrasound catheter probes will be equipped with an integrated electromagnetic tracking sensor. To benefit from the advantages of freehand 3D ultrasound the catheters have to be calibrated. However most of the methods proposed in the literature have been developed for prototype work. For the calibration of the catheters during their fabrication there is a need for a fast, efficient, and fully automatic calibration process.

In this work I have developed a framework for the fully automatic calibration of ultrasound catheters with an integrated electromagnetic tracking sensor. Within the framework two calibration phantoms were developed. The first phantom uses spheres as markers, the second phantom uses modulated circles as markers. For both phantom types I have implemented a fully automatic calibration process. The developed methods satisfy the strict requirements for an industrial calibration process. The results are evaluated by studies on synthetic ground truth data sets and experiments with real ultrasound images.

Resources

Literature