VidSfM: Robust and Accurate Structure-From-Motion for Monocular Videos.
Hainan CuiDiantao TuFulin TangPengfei XuHongmin LiuShuhan ShenPublished in: IEEE Trans. Image Process. (2022)
Keyphrases
- structure from motion
- image sequences
- bundle adjustment
- camera pose estimation
- depth estimates
- video sequences
- point cloud
- camera parameters
- missing data
- camera motion
- computer vision
- scene structure
- reconstruction from multiple images
- moving camera
- focal length
- reprojection error
- motion analysis
- pose estimation
- motion estimation
- feature tracking
- camera calibration
- machine learning
- photometric information
- motion estimates
- feature correspondences
- rigid motion
- image noise
- optical flow
- ego motion
- visual slam
- motion parameters
- background subtraction