Scene Text Image Super-resolution based on Text-conditional Diffusion Models.
Chihiro NoguchiShun FukudaMasao YamanakaPublished in: CoRR (2023)
Keyphrases
- image super resolution
- scene text
- diffusion models
- text detection
- natural scene images
- super resolution
- diffusion model
- low resolution
- dictionary learning
- information diffusion
- text regions
- image patches
- video frames
- image registration
- maximum a posteriori
- text lines
- sparse representation
- social networks
- complex background
- connected components
- text information
- document images
- object detection
- input image
- feature extraction