A learned sparseness and IGMRF-based regularization framework for dense disparity estimation using unsupervised feature learning.
Sonam NaharManjunath V. JoshiPublished in: IPSJ Trans. Comput. Vis. Appl. (2017)
Keyphrases
- gaussian markov random field
- regularization framework
- autoregressive
- semi supervised learning
- edge preserving
- super resolution
- stereo images
- metric learning
- weighted graph
- semi supervised
- energy function
- labeled and unlabeled data
- unlabeled data
- stereo vision
- disparity map
- map estimation
- image pairs
- object recognition
- low resolution
- least squares
- high resolution