A novel structure-from-motion strategy for refining depth map estimation and multi-view synthesis in 3DTV.
Chia-Ming ChengXiao-An HsuShang-Hong LaiPublished in: ICME (2010)
Keyphrases
- view synthesis
- structure from motion
- scene reconstruction
- depth map
- image sequences
- multi view
- bundle adjustment
- d scene
- image based rendering
- depth information
- camera parameters
- multiple cameras
- stereo matching
- photorealistic
- point cloud
- missing data
- computer vision
- camera motion
- stereo pair
- depth estimation
- camera calibration
- stereo correspondence
- high quality
- three dimensional
- multiple views
- focal length
- surface reconstruction
- object tracking
- rigid motion
- input image