Laparoscopic Feature-Less 3D Reconstruction Using Neural Radiance Fields and Optical Tracking

The use of laparoscopic liver or pancreatic surgery is increasing compared to open surgery, but it presents challenges like limited field of view and lack of tactile feedback. Mentally correlating the intra-operative scene with pre-operative volumes is challenging. Therefore, surgical navigation systems aim to assist with intraoperative understanding, and one approach is performing intraoperative 3D scene reconstructions. Neural Radiance Fields (NeRF) have the potential to photorealistically reconstruct features-less scenes if the poses of the camera are accurately known. This study investigates the accuracy of 3D reconstructions combining NeRF with optical tracking. The accuracy of the NeRF 3D reconstruction is assessed on an abdominal phantom and corresponding CT. Results show that the combination of optical tracking and NeRF shows promise in meeting the clinical navigation accuracy requirements for laparoscopic surgery.