StructDepth: Leveraging the structural regularities for self-supervised indoor depth estimation.
Boying LiYuan HuangZeyu LiuDanping ZouWenxian YuPublished in: ICCV (2021)
Keyphrases
- depth estimation
- stereo vision
- depth map
- stereo matching
- scene understanding
- depth information
- dynamic scenes
- super resolution
- model based pose estimation
- stereo pair
- real scenes
- feature matching
- multi view
- viewpoint
- preprocessing
- computer vision
- stereo images
- high quality
- real time
- multi class
- depth from defocus
- data sets
- three dimensional
- image sequences
- object boundaries
- pairwise
- d scene
- d objects
- vision system