Devanagari Character Dataset

Context

This dataset is created as a part of my dissertation work for the fulfillment of the Master's degree in Computer Science (Tribhuvan University, Nepal, 2012).

Content

The dataset contains three individual categories. Samples are collected from 40 individuals (persons) from different fields and cropped for character boundary.

1. Numerals (288 samples per class, 10 classes)
2. Vowels (221 samples per class, 12 classes)
3. Consonants (205 samples per class, 36 classes)

Citation

Please cite in your publications if it helps your research:

@inproceedings{pant2012off,
  title={Off-line Nepali handwritten character recognition using Multilayer Perceptron and Radial Basis Function neural networks},
  author={Pant, Ashok Kumar and Panday, Sanjeeb Prasad and Joshi, Shashidhar Ram},
  booktitle={2012 Third Asian Himalayas International Conference on Internet},
  pages={1--5},
  year={2012},
  organization={IEEE}
}