Alphabet Recognizer

Simple Optical Character Recognizer

This Image processing program can convert an image of English writing (like- this) to text. It simply uses the Template Matching strategy leveraged by Bayes' theorem for character recognition.

How to Run?

1. Download Alphabet Recognizer, and import it in your IDE.
2. Download egami.jar from project- Egami, and integrate it in the project's build path.
3. For training the machine run Train.java.
4. For testing run Predict.java (change file-paths in the main method according to your need).

Visual of Learned Templates

<div align="center"> <img src="https://user-images.githubusercontent.com/5456665/27283437-1a69706c-5516-11e7-9410-cf169d325a43.png" height="500" width=auto title="Template Knowledge"> </div>

Downloads

• AlphabetRecognizer (zip)
• English Alphabet Dataset (zip)

If you like this project, you may also like GenderRecognizer- Plain Face Detector & Gender Recognizer and SkinDetector- Detects Human Skin From Image.

Citation

@misc{Minhas2016AlphabetRecognizer,
  author    = {Kamal, Minhas},
  title     = {Alphabet Recognizer: Simple Optical Character Recognizer},
  year      = {2015},
  publisher = {GitHub},
  url       = {https://github.com/MinhasKamal/AlphabetRecognizer}
}

License

<a rel="license" href="https://opensource.org/licenses/MIT"><img alt="MIT License" src="https://cloud.githubusercontent.com/assets/5456665/18950087/fbe0681a-865f-11e6-9552-e59d038d5913.png" width="60em" height=auto/></a><br/><a href="https://github.com/MinhasKamal/AlphabetRecognizer">Alphabet Recognizer</a> project and its <a href="https://github.com/MinhasKamal/AlphabetRecognizer/tree/master/src/res/trainingData">Training Dataset</a> is licensed under <a rel="license" href="https://opensource.org/licenses/MIT">MIT License</a>.