MVSFormer++: Revealing the Devil in Transformer's Details for Multi-View Stereo.
Chenjie CaoXinlin RenYanwei FuPublished in: ICLR (2024)
Keyphrases
- multi view stereo
- multi view
- structure from motion
- depth map
- stereo reconstruction
- bundle adjustment
- urban environments
- point cloud
- high fidelity
- camera parameters
- ray tracing
- depth estimation
- surface reconstruction
- camera calibration
- scene reconstruction
- real time
- visual hull
- multiple views
- computer vision
- depth information
- complex objects
- image synthesis
- scene structure
- pairwise
- high quality
- stereo algorithms