There is discrepancy between the exact representation of anatomical structures and tumours in the CT or MRI scan and the more or less accurate intra-operative localisation methods based mostly upon landmarks of the skull and extracerebral space and visible abnormalities of the cerebral surface. To overcome these problems of exact intra-operative localisation a Computer Assisted Localizer (CAL) is presented which allows precise intra-operative orientation without these aids. It consists of a mechanically articulated arm with six degrees of freedom with a high precision digital incremental and an image processor for 3 D data of the head. MRI and/or CT investigation is done pre-operatively with four reference markers fixed on the patient's head. They are visible on the CT or MRI slices and are used as reference points during surgery for adjustment of the device. The co-ordinates of the digitalizer arm tip are projected into the corresponding axial, sagittal and coronal CT slices so that the system simultaneously presents three orthogonal multiplanar CT reconstructions with a reticule indicating the position of the tip of the arm. As the surgeon directs the arm to the region of interest the corresponding CT slices are displayed on the monitor at a rate of 20 slices/sec determined by the motion of the arm. The accuracy of measurement of the device itself lies within 1 mm. The accuracy is somewhat reduced however by the thickness of CT or MRI slices (routinely 2 mm slices were taken) and by deviations of the reference markers on the skin surface which amount up to 3 mm.(ABSTRACT TRUNCATED AT 250 WORDS)

No abstract

No abstract

The three-dimensional reconstruction of rigid scenes from monocular image streams is based on the former calculation of the relative camera pose between at least two successive image frames. This egomotion estimation has not been solved satisfactorily by relying only on corresponding image features, such as points or lines, due to noise, critical motion patterns or special point configurations. This paper describes a framework for incorporating inertial measurements from gyroscopes, accelerometers and magnetometers to achieve an improved performance of the estimation of camera motion and scene structure in terms of accuracy, robustness and computational costs. The framework is designed as a dual-track system containing a visual and a inertial route.

No abstract