Abstract #2947

Robust Computer Assisted Catheter Tracking Algorithm in MR Using Gradient Based Signatures and Mean Shift Localization

Abubakr El-Tahir¹,², Frederic Courivaud¹, Rafael Palomar¹, Ole Jakob Elle¹,²

¹The Intervention Centre, Oslo University Hospital, Oslo, Norway; ²Department of Informatics, University of Oslo, Oslo, Norway

A smart robust algorithm for automatic catheter tracking in MR is proposed. The output of the catheter tracking is intended to be used in developing an automatic MR image acquisition and visualization scheme to be presented to the interventionist during catheterization procedures. In this method the catheter tip is marked with an MR traceable marker (in this case a miniature resonant circuit). Then automatic tip tracking is performed in a series of MR frames using probability map generated for each frame based on the similarity to a pre-learned template marker.