Synthetic Data Generation for Arabic Scene Text Recognition

  This is the synthetic Arabic data generator from the paper "Arabic Scene Text Recognition in the Deep Learning Era: Analysis on a Novel Dataset"

This data generator is based on the generator from the paper "Editing Text in the Wild"  

Arabic Fonts

Background pictures

 

Requirements

• Python 3.6
• numpy
• opencv-python
• imgaug
• bidi
• arabic_reshaper

References

• SRNet-Datagen - A data generator of SRNet

• Synthetic Data for Text Localisation in Natural Images