This work adapted the method of Wei-Sheng Lai et al. to do super-resolution on (English, French and Vietnamese) text images. We also added Gradient Different Loss (GDL) to L1 loss to sharpen the output.
Training and testing datasets can be downloaded here.