On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach.
Nikolai SmolyanskiyAlexey KamenevStan BirchfieldPublished in: CVPR Workshops (2018)
Keyphrases
- depth estimation
- semi supervised
- depth map
- stereo vision
- depth estimates
- binocular stereo
- stereo matching
- multi view
- motion parallax
- stereo pair
- depth information
- dynamic scenes
- high quality
- scene understanding
- real scenes
- depth perception
- stereo images
- disparity map
- super resolution
- dense depth maps
- range sensors
- object recognition
- feature matching
- blur estimation
- d scene
- belief propagation
- structure from motion
- specular reflection
- multiple views
- input image
- depth from defocus
- video sequences
- three dimensional
- learning algorithm