MAML-SR: Self-adaptive super-resolution networks via multi-scale optimized attention-aware meta-learning.
Debabrata PalShirsha BoseDeeptej MoreAnkit JhaBiplab BanerjeeYogananda V. JeppuPublished in: Pattern Recognit. Lett. (2023)
Keyphrases
- super resolution
- meta learning
- multiscale
- low resolution
- high resolution
- learning tasks
- inductive learning
- image reconstruction
- model selection
- low resolution images
- high quality
- resolution enhancement
- motion estimation
- image super resolution
- single image super resolution
- super resolution reconstruction
- depth map
- decision trees
- multi frame
- machine learning algorithms
- image patches
- super resolved
- video super resolution
- feature selection
- machine learning
- high resolution images
- base classifiers
- video sequences
- data mining
- training samples