EnPrO: Enhancing Precision Through Optimization in Image-Guided Spine Surgical Procedures

Advancements in intra-operative visualization have accelerated the adoption of C-Arm Fluoroscopy imaging modalities within Image-Guided Spine Surgery (IGSS) procedures. The proposed research provides a novel technique for improving precision in IGSS via EnPrO by refining the mapping of 2D fluoroscopic images to the patient’s anatomy. The fundamental strategy is to minimize reprojection error (RPE) by picking optimal fiducial sites via weighted norm approximation. In EnPrO, we propose two methods to perform optimization for fiducial weights, namely, using fiducial coordinates and using a camera projection matrix (CPM). Using EnPrO, the C-Arm imaging distortion that contributes to increasing RPE can be identified and excluded from the IGSS calibration procedure. Using EnPrO, fiducials were chosen based on weights obtained using the fiducial coordinates and the CPM clearly showed an average RPE decrease of 6.96% and 8.36% respectively. The implementation of the project can be found in this repository: