https://github.com/masich/simple-number-recognizer

Simple number recognizer created using pure Python 3
https://github.com/masich/simple-number-recognizer

neural-network number-recognition perceptron perceptron-learning-algorithm python

Last synced: 3 months ago
JSON representation

Simple number recognizer created using pure Python 3

• Host: GitHub
• URL: https://github.com/masich/simple-number-recognizer
• Owner: masich
• License: mit
• Created: 2018-11-07T23:51:34.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2023-02-05T23:27:42.000Z (over 2 years ago)
• Last Synced: 2025-01-27T19:38:32.774Z (5 months ago)
• Topics: neural-network, number-recognition, perceptron, perceptron-learning-algorithm, python
• Language: Python
• Homepage:
• Size: 282 KB
• Stars: 2
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# SimpleNumberRecognizer

### Perceptron-based simple number recognizer written using pure Python 3.

[![CodeFactor](https://www.codefactor.io/repository/github/masich/simple-number-recognizer/badge)](https://www.codefactor.io/repository/github/masich/simple-number-recognizer)

[![License](https://img.shields.io/badge/License-MIT-red.svg)](LICENSE)

This project is created just for fun and practice.

The project's goal is very primitive: the user "draws" some numbers (0-9) using buttons, and the program tries to guess the drawn number. You can teach an internal perceptron-based neural network or use a simple predefined learning dataset in a ```data``` folder. 

## Instruction

![Interface](images/readme/interface.png)

## For teaching

* Use buttons **(1)** to draw the number you want to teach the system.

* Give the name of the number class (0 - 9) by putting it into the Class name field **(4)**.

* Click the Done button **(5)**. Now your number and its class are saved into internal program memory.

* Repeat these two steps as often as you wish to teach the system as best as possible.

* Click the Save file **(8)**. All your inputted numbers and classes will be saved as ```data/learning_data.json```.

* Click on Learn button **(9)** to teach your system using ```data/learning_data.json```.

## For testing

* Click on Learn button **(9)** to teach your system using ```data/learning_data.json```.

* Use buttons **(1)** to draw the number.

* Click Test button **(2)** to try to recognize the number.

* The output will be displayed on the **(3)** field. Each perceptron corresponds to a particular number. For example, 

the system thinks that you have drawn the number 4 if the value of the perceptron 4 is ```True```.   

* To clear the input **(1)** and output **(3)** you can use Clear button **(7)**.

## Screenshots

![Example_1](images/readme/example_1.png)

![Example_2](images/readme/example_2.png)

![Example_3](images/readme/example_3.png)

## License

This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.