Generative Adversarial CT Volume Extrapolation for Robust Small-to-Large Field of View Registration

Intraoperative Computer Tomographs (iCT) provide near real time visualizations which can be registered with high-quality preoperative images to improve the confidence of surgical instrument navigation. However, intraoperative have a small field view making registration process error prone due reduced amount mutual information. We herein propose method extrapolate thin acquisitions as prior step registration, increase images, and hence also robustness guiding system. The is based on deep neural network trained adversarially using self-supervision slices from existing ones. Median landmark detection errors are by approximately 40%, yielding better initial alignment. Furthermore, intensity-based improved; surface distance an order magnitude, 5.66 mm 0.57 (p-value = 4.18×10−6). proposed extrapolation increases robustness, plays key role in intervention confidently.