CT-SRCNN: Cascade Trained and Trimmed Deep Convolutional Neural Networks for Image Super Resolution.
Haoyu RenMostafa El-KhamyJungwon LeePublished in: CoRR (2017)
Keyphrases
- convolutional network
- convolutional neural networks
- image super resolution
- dictionary learning
- super resolution
- low resolution
- image registration
- maximum a posteriori
- high resolution
- coarse to fine
- edge preserving
- low resolution images
- sparse representation
- kernel regression
- image patches
- map estimation
- denoising
- sparse coding
- em algorithm
- expectation maximization