This dataset has 60,122 binary TIF text line images across different Sindhi fonts. Ground truth provided as UTF-8 encoded .txt files and .xml metadata files. Train/test splits provided as split files for reproducibility.
For academic and non-commercial research use only.