-
1
X-ray imaging is essential for orthopedic procedures, aiding in diagnosis, intraoperative guidance, and surgical planning.
-
2
2D/3D registration reconstructs patient-specific 3D models from 2D radiographs, improving surgical planning while minimizing radiation exposure.
-
3
The proposed deep learning network estimates a registration field from calibrated radiographs, avoiding the need for additional segmentation.
-
4
The registration method decomposes the transformation into affine and local components, enhancing flexibility in input data orientation.
-
5
Validation of the network was performed using simulated digitally reconstructed radiographs, demonstrating its effectiveness compared to existing methods.
