sahil485 / CharacterRecog

One of my first personal ML projects using Python's Tensorflow Keras and Tkinter libraries to create canvas for users to draw a number on with a mouse. Then, a neural network trained on the MNIST handwritten numbers dataset is used to classify the digit.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

DigitRecog

One of my first personal ML projects using Python's Tensorflow Keras and Tkinter libraries to create canvas for users to draw a number on with a mouse. Then, a neural network training on the MNIST handwritten numbers dataset is used to classify the digit.

References:

https://github.com/machinecurve/extra_keras_datasets - https://github.com/ghanteyyy/tkcap

Steps to Play:

  1. Clone repo
  2. Run the DrawingBoard.py file
  3. While hovering with your mouse over the tkinter window left-click once and draw a digit in one "motion"
  4. Left click again once finished
  5. If the machine's classification is incorrect, click on the header of the tkinter window and type 'x'
  6. Right click while the mouse is hovering over the tkinter window at any time to see an accuracy report of the session and exit

About

One of my first personal ML projects using Python's Tensorflow Keras and Tkinter libraries to create canvas for users to draw a number on with a mouse. Then, a neural network trained on the MNIST handwritten numbers dataset is used to classify the digit.

License:MIT License


Languages

Language:PureBasic 73.6%Language:Python 26.4%