Is my Depth Ground-Truth Good Enough? HAMMER - Highly Accurate Multi-Modal Dataset for DEnse 3D Scene Regression.
HyunJun JungPatrick RuhkampGuangyao ZhaiNikolas BraschYitong LiYannick VerdieJifei SongYiren ZhouAnil ArmaganSlobodan IlicAles LeonardisBenjamin BusamPublished in: CoRR (2022)
Keyphrases
- multi modal
- highly accurate
- d scene
- depth map
- high quality
- ground truth
- scene flow
- view dependent
- depth information
- view synthesis
- scene flow estimation
- scene geometry
- flow estimation
- image based rendering
- photorealistic
- scene understanding
- camera parameters
- ego motion
- multi view
- depth estimation
- high dimensional
- stereo matching
- high accuracy
- single image
- scene structure
- depth data
- depth images
- optical flow
- super resolution
- cross modal
- camera viewpoint
- uni modal
- image sequences
- single modality
- planar patches
- range scans
- viewpoint
- high resolution
- active learning
- input image
- input data
- three dimensional