Intraoperative zoom lens calibration for high magnification surgical microscope

Background and objectives: An augmented reality (AR)-based surgical guidance system is often used with high-magnification zoom lens systems such as a surgical microscope, particularly in neurology or otolaryngology. To superimpose the internal structures of relevant organs on the microscopy image, an accurate calibration process to obtain the camera intrinsic and hand–eye parameters of the microscope is essential. However, conventional calibration methods are unsuitable for surgical microscopes because of their narrow depth of focus at high magnifications. To realize AR-based surgical guidance with a high-magnification surgical microscope, we herein propose a new calibration method that is applicable to the highest magnification levels as well as low magnifications. Methods: The key idea of the proposed method is to find the relationship between the focal length and the hand–eye parameters, which remains constant regardless of the magnification level. Based on this, even if the magnification changes arbitrarily during surgery, the intrinsic and hand–eye parameters are recalculated quickly and accurately with one or two pictures of the pattern. We also developed a dedicated calibration tool with a prism to take focused pattern images without interfering with the surgery. Results: The proposed calibration method ensured an AR error of < 1 mm for all magnification levels. In addition, the variation of focal length was within 1% regardless of the magnification level, and the corresponding variation with the conventional calibration method exceeded 20% at high magnification levels. Conclusions: The comparative study showed that the proposed method has outstanding accuracy and reproducibility for a high-magnification surgical microscope. The proposed calibration method is applicable to various endoscope or microscope systems with zoom lens.