Bi-ESRGAN: A New Approach of Document Image Super-Resolution Based on Dual Deep Transfer Learning.
Zakia KezzoulaDjamel GacebZineddine AkliAbdelouaheb KahouliAyoub TitounFayçal TouaziPublished in: ICAITA (2022)
Keyphrases
- transfer learning
- image super resolution
- low resolution
- super resolution
- labeled data
- dictionary learning
- reinforcement learning
- active learning
- machine learning
- transfer knowledge
- cross domain
- semi supervised learning
- image registration
- information retrieval
- edge preserving
- machine learning algorithms
- high resolution
- target domain
- collaborative filtering
- multi task
- kernel regression
- text categorization
- learning algorithm
- maximum a posteriori
- low resolution images
- text mining
- image patches
- high quality
- sparse coding
- computer vision
- feature extraction
- training examples
- image restoration
- semi supervised
- text classification
- data sets