File size: 1,096 Bytes
c9828df d2a8168 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 |
# Kashmiri OCR Dataset
## Dataset Overview
This dataset is designed for Optical Character Recognition (OCR) tasks involving the Kashmiri language. It contains images generated with different fonts and texts from a Kashmiri text corpus.
## Dataset Structure
- **Training Set**: 85% of the total samples.
- **Validation Set**: 5% of the total samples.
- **Test Set**: 10% of the total samples.
## File Structure
- `train/`: Directory containing training images and `labels.csv`.
- `validation/`: Directory containing validation images and `labels.csv`.
- `test/`: Directory containing test images and `labels.csv`.
- `combined_files/`: Directory containing images and their respective text files (optional).
- `metadata.json`: Metadata file with details about dataset generation.
- `metadata.yaml`: Metadata file in YAML format with details about dataset generation.
## Usage
The dataset is divided into training, validation, and test sets. You can use the CSV files for loading image-text pairs.
## License
This dataset is provided for educational purposes.
## Author
Haq Nawaz Malik |