A learned overcomplete sparseness and IGMRF based regularization framework for dense disparity estimation.
Sonam NaharManjunath V. JoshiPublished in: ACPR (2015)
Keyphrases
- gaussian markov random field
- disparity estimation
- regularization framework
- autoregressive
- super resolution
- motion estimation
- stereo images
- edge preserving
- semi supervised learning
- semi supervised
- vision system
- gray scale
- weighted graph
- metric learning
- image pairs
- bilateral filter
- depth information
- natural images
- sparse representation
- stereo matching
- image restoration
- map estimation
- disparity map
- depth images
- energy function
- labeled and unlabeled data
- depth map
- training data
- anisotropic diffusion
- motion vectors
- machine learning
- markov random field
- active learning
- high resolution
- multiresolution
- video sequences
- image sequences