A Projection-based AR system to display brain angiography via Stereo Vision

In this paper, a projector-based augmented reality (AR) system to display segmented brain angiography is proposed. Firstly, this scheme employed an adaptive algorithm to find the optimal threshold to segment the computer tomography's (CT) vascular image. Then, with the help of computer vision technology, the geometrical relationship between the facial data and the CT's angiography information is derived, and the registration process is executed. When the image registration is finished, the correct location of the vascular image is obtained. Since the surface of the head is irregular, directly project the uncorrected CT angiography images on the surface will be deformed. To alleviate this problem, we employed a projector to project a set of lines on the head, and then a camera is used to capture these lines to derive the correction information required for surface deformation. Finally, the vascular image after deformation correction is projected on the head to provide the doctors as a reference.