Search Results for author:
Found 3 papers, 2 papers with code
Visual Geometry Grounded Deep Structure From Motion
Finally, we optimise the cameras and triangulate 3D points via a differentiable bundle adjustment layer.
CoTracker: It is Better to Track Together
We introduce CoTracker, a transformer-based model that tracks dense points in a frame jointly across a video sequence.
DynamicStereo: Consistent Dynamic Depth from Stereo Videos
The network learns to pool information from neighboring frames to improve the temporal consistency of its predictions.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
