Grant: Alliance Grants

Geometric camera calibration is a fundamental part of any vision system that uses an optical sensor. It defines a camera model representing the projective transformation between the 3D world and the 2D image plane. Due to the projective nature of the transformation, minor calibration errors translate into significant errors in the world coordinate space, which can have a detrimental effect on the vision system's accuracy and performance. The work in this proposal aims to develop solutions that improve the accuracy of the calibration of a multi-camera system that is part of an active robotic surgical system for orthopaedic surgery developed, manufactured and marketed by our industry partner, THINK Surgical Inc. Based on preliminary investigations carried out by THINK Surgical Inc's researchers the proposed research methodology is to develop new algorithms for (i) calibrating the multi-camera system, (ii) detecting and tracking of markers and surgical instruments in the presence of occlusions, (iii) reconstructing in 3D the location of the surgical instruments, and to develop new measures and metrics appropriate for the (iv) evaluation of each component.The project will facilitate the training of 3 HQPs (1 Ph.D., 2 Masters, 1 USRA) in the cutting edge topic of vision-guided robotic surgery, which is of great interest and benefit to the Canadian scientific and industrial communities. The fundamental research and investigations carried out as part of this project will open new research directions in computer vision and its use within the medical context. The results will benefit Canada, as we anticipate that THINK Surgical Inc will use them to provide more accurate, faster and globally competitive solutions to its clients worldwide.