Sensor fusion for surgical applications

Abstract

The CAMIS prototype surgical navigation system combines preoperative 3D imagery and intra-operative localization to register a patient with this imagery. While this is vastly superior to the exploratory surgery that it replaced, improving anatomical localization is one of the primary goals of the CAMIS project. This presentation discusses the fusion of preoperative and intraoperative ultrasound mensuration imagery. In particular, a PC based ultrasound system developed at Washington University was modified to add localization capabilities. Then using a unique single point calibration procedure, the projection of the ultrasound 2 dimensional coordinates into full 3 dimensional patient coordinates was developed. Thus the 3D coordinates of any point in the ultrasound image can be directly calculated from this mapping. The fusion of preoperative CT imagery with this calibrated real time ultrasound is currently being tested using a spinal phantom. A 3D CT image set of the phantom was acquired and is used as a 3D model for generating a synthetic ultrasound axial image slice. This synthetic ultrasound image was generated using the Ohio University Synthetic Ultrasound code developed under the CAMIS project. Then the spinal phantom was imaged while submerged in water and the acquisition geometry was manually aligned to get the best apparent match between the measured and synthetic ultrasound images. These two images were then computer matched by adjusting the synthetic acquisition parameters until a best match was obtained. A coordinate transformation is then shown that convert from CT coordinates to patient coordinates so that the patient and the CT image are geometrically registered through the intra-operative ultrasound image.