Registration of 3D Angiographic and X-Ray Images Using Sequential Monte Carlo Sampling

Abstract. Digital subtraction angiography (DSA) reconstructions and 3D Magnetic Resonance Angiography (MRA) are the modalities of choice for diagnosis of vascular diseases. However, when it comes to treatment through an endovascular intervention, only two dimensional lower resolution information such as angiograms or fluoroscopic images are usually available. Overlaying the preoperative information from high resoluion acquisition onto the images acquired during intervention greatly helps physician in performing the operation. We propose to register pre-operative DSA or MRS with intra-operative images to bring the two data sets into a single coordinate frame. The method uses the vascular structure, which is present and visible from most of DSA, MRA and x-ray angiogram and fluoroscopic images, to determine the registration parameters. A robust multiple hypothesis framework is built to minimize a fitness measure between the 3D volume and the 2D projection. The measure is based on the d...

• Type: Conference
• Year: 2005
• Where: CVBIA
• Authors: Charles Florin, James Williams, Ali Khamene, Nikos Paragios
• Keywords: Computer Vision, CVBIA 2005, Digital Subtraction Angiography, Magnetic Resonance Angiography, fluoroscopic Images