https://github.com/orangomango/handwrittendigitsrecognizer

Handwritten digits recognizer by using a neural network
https://github.com/orangomango/handwrittendigitsrecognizer

ai java javafx neural-network

Last synced: about 1 month ago
JSON representation

Handwritten digits recognizer by using a neural network

• Host: GitHub
• URL: https://github.com/orangomango/handwrittendigitsrecognizer
• Owner: OrangoMango
• License: mit
• Created: 2024-06-16T08:38:47.000Z (about 2 years ago)
• Default Branch: main
• Last Pushed: 2024-06-16T08:59:56.000Z (about 2 years ago)
• Last Synced: 2025-02-26T15:38:00.177Z (over 1 year ago)
• Topics: ai, java, javafx, neural-network
• Language: Java
• Homepage: https://youtube.com/shorts/ZE2NT-mdVBc?si=nZa1vGP0JAuYuont
• Size: 220 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# HandwrittenDigitsRecognizer

Handwritten digits recognizer by using a neural network.  

Short video about it: [YouTube Shorts](https://youtube.com/shorts/ZE2NT-mdVBc?si=nZa1vGP0JAuYuont)

# How to use

* Download the MNIST training images from [here](http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz)

* Download the MNIST training labels from [here](http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz)

* Compile the neural network file: `javac NeuralNetwork.java`

* Compile the JavaFX file: `javac --module-path $FX_PATH --add-modules javafx.controls HandRecognizer.java` were `$FX_PATH` contains the path of your lib folder in the javafx installation directory.

* (optional) Train the model: `java NeuralNetwork.java`

* Run the JavaFX application: `java --module-path $FX_PATH --add-modules javafx.controls HandRecognizer`

# Instructions

* Draw a number on the left (left mouse button) and the network will tell you what number it is and the confidence.

* Use the `DELETE` key to clear the canvas

* Use the digits from `0` to `9` to train the network while using it (for example if it recognizes some digits wrongly)

* Use the `L` key to load a random image of the database

All the training data is saved in the file `training.txt`

# Screenshot

![Screenshot from 2024-06-16 10-48-39](https://github.com/OrangoMango/HandwrittenDigitsRecognizer/assets/61402409/4866073f-8b18-4dd2-bc1c-e757dcd15a07)