/TextGenerator

OCR dataset Text-Detection dataset Font-Classification dataset generator

Primary LanguagePythonMIT LicenseMIT

TextGenerator

  • This is a tool for ocr dataset, text detection, fonts classification dataset generate.

Functions：

  • different font, size, color, rotation
  • multithreading
  • specific layout
  • find smooth area to attach
  • character-wised annotation

image generated:

image attach:

rotation

bounding box of single word

```
# step 1
pip install requirements.txt
# step 2
sh make.sh
```

  • config file config.yml

  • run script

    python3 run.py

  • data generated config.yml -> provider> layout> out_put_dir

content modified

  1. np.int related
  2. check the piplist.txt for package version

todo:

Math generation and manipulation for formula detection task