Fast 3D Structure From Motion with Missing Points from Registration of Partial Reconstructions

The status of augmented reality in laparoscopic surgery as of 2016

“…Based on a more sparse approach and relying on rigid points in the scene, SLAM is less computationally demanding and more robust than SfM, but requires a good initialization and the rigidity of most the scene. Variations of SLAM include Extended-Kalman-Filter-SLAM (EKF-SLAM), Parallel Tracking And Mapping (PTAM) ( Grasa et al, 2014 ), SfM/SLAM hybrid ( Mahmoud et al, 2012;Collins et al, 2013 ) and Motion-compensated SLAM ( Mountney and Yang, 2010 ). 2014) .…”

Optical techniques for 3D surface reconstruction in computer-assisted laparoscopic surgery

“…Camera localization is performed using whole image alignment with the dense models and does not rely on feature matching. In current research, there is a trend towards online SFM and hybrid SFM/SLAM systems which combine the benefits of local batch processing with online localization (Klein and Murray, 2007;Newcombe et al, 2011;Mahmoud et al, 2012). Further active research areas include loop closing, dynamic environments and long-term mapping.…”

Image-guided surgery

“…220 Mountney and colleagues, 224 Stoyanov and colleagues, 223 Lee and colleagues, 222 and Röhl and colleagues 221 typically used a stereoscopic endoscope to reconstruct organ surface using stereovision technique (needing a calibration step before the surgery begins). Mountney and colleagues, 227 Hu and colleagues, 225 Mahmoud and colleagues, 228 and Garcia and colleagues 226 proposed a solution to estimate the organ surface from the camera motion with simultaneous localization and mapping or structure from motion technique, the main idea being that parallax motion is sufficient to understand the scene shape. Other teams also use image information only, shading, 229,230 and shape assumption 231,232 (using conformal model), to estimate the organ surface with satisfactory results.…”