yashpaneliya / handwritten-digit-recognition

Handwritten-Digit-Recognition

Convolutional Neural Network trained on 60000 images of handwritten digits

Installation setup

• install the following libraries for training purposes

  • numpy
  • keras
  • matplotlib
  • sklearn

• install the following libraries for testing the model

  • tkinter
  • keras
  • CV2
  • numpy
  • PIL

• Open handwritten_digits1.ipynb in Jupyter notebook or Google colab

• execute all the cells consecutively to train the model

• model will be saved at the end with name 'digitdemo1.h5'

• run scatchboard.py to test the model by drawing some digits on canvas