Innovative Approches to 2D/3D Registration for Image-Guided Therapy

2D/3D registration is a vital field of research in applied medical image processing and medical physics. In short, 2D/3D registration deals with the automated transfer of preoperatively acquired volume images from computed tomography or magnetic resonance imaging to a coordinate system associated with the patient. Applications include radiation therapy, telemanipulator-assisted surgery, or the field of image-guided therapy and interventional radiology. In this proposal, novel strategies for 2D/3D intensity based registration in image guided therapy are introduced. These strategies include novel merit functions for image comparison based on rotation-invariant Zernike expansions, the evaluation of registration merit functions and innovative rendering algorithms for generation of digitally rendered radiographs on common graphics hardware using mobile medium - sized computer clusters, and advanced optimization algorithms. A comprehensive research effort into these new paradigms will result in development of new, faster and more robust 2D/3D registration algorithms, thereby paving the way to novel clinical applications like slice-to-volume registration and biomechanical analysis. The long-term goal of our effort is non-rigid 2D/3D registration for support in telemanipulator - assisted cardiac surgery. Furthermore, we will perform clinical validation studies for this application. A tangible impact of the research project on the field of medical image processing and on society in general is expected. More reliable and faster registration methods will advance the patient - friendly way of non-invasive registration to a wide number of applications beyond the aforementioned fields.. As a general outcome we expect improved patient benefit and reduced hospitalization in fields like tumor biopsy or cardiac surgery.

2D/3D registration is a vital field of research in applied medical image processing and medical physics. In short, 2D/3D registration deals with the automated transfer of preoperatively acquired volume images from computed tomography or magnetic resonance imaging to a coordinate system associated with the patient. Applications include radiation therapy, telemanipulator-assisted surgery, or the field of image-guided therapy and interventional radiology. In this proposal, novel strategies for 2D/3D intensity based registration in image guided therapy are introduced. These strategies include novel merit functions for image comparison based on rotation-invariant Zernike expansions, the evaluation of registration merit functions and innovative rendering algorithms for generation of digitally rendered radiographs on common graphics hardware using mobile medium - sized computer clusters, and advanced optimization algorithms. A comprehensive research effort into these new paradigms will result in development of new, faster and more robust 2D/3D registration algorithms, thereby paving the way to novel clinical applications like slice-to-volume registration and biomechanical analysis. The long-term goal of our effort is non-rigid 2D/3D registration for support in telemanipulator - assisted cardiac surgery. Furthermore, we will perform clinical validation studies for this application. A tangible impact of the research project on the field of medical image processing and on society in general is expected. More reliable and faster registration methods will advance the patient - friendly way of non-invasive registration to a wide number of applications beyond the aforementioned fields.. As a general outcome we expect improved patient benefit and reduced hospitalization in fields like tumor biopsy or cardiac surgery.