3D Scene Mesh from CNN Depth Predictions and Sparse Monocular SLAM.
Tomoyuki MukasaJiu XuBjörn StengerPublished in: ICCV Workshops (2017)
Keyphrases
- d scene
- depth map
- monocular slam
- view synthesis
- depth information
- multi view
- bundle adjustment
- camera parameters
- single image
- visual slam
- single camera
- scene structure
- ego motion
- depth estimation
- scene understanding
- scene reconstruction
- high quality
- moving camera
- image based rendering
- stereo matching
- image sequences
- three dimensional
- photorealistic
- d mesh
- optical flow
- input image
- real scenes
- depth images
- planar surfaces
- high resolution
- stereo vision
- pairwise
- stereo pair
- multiple views
- visual tracking
- camera positions
- sparse representation