Depth Estimation Through a Generative Model of Light Field Synthesis.
Mehdi S. M. SajjadiRolf KöhlerBernhard SchölkopfMichael HirschPublished in: CoRR (2016)
Keyphrases
- generative model
- light field
- depth estimation
- super resolution
- depth map
- multi view
- stereo vision
- dynamic scenes
- low resolution
- probabilistic model
- high resolution
- semi supervised
- stereo matching
- image formation
- depth information
- em algorithm
- real scenes
- single image
- image restoration
- prior knowledge
- motion estimation
- stereo pair
- high quality
- motion blur
- view synthesis
- d scene
- d objects
- three dimensional
- image processing
- expectation maximization
- graphical models
- video camera
- multiple cameras
- face images
- feature matching
- dynamic programming
- active learning
- object recognition