/handwritten-MNIST-digit-recognition

Real time MNIST digit recognition with OpenCV and Support Vector Machine (SVM) algorithm.

Primary LanguageJupyter NotebookMIT LicenseMIT

Handwritten Digit Recognition Using OpenCV and Python

Dataset used

For this project I used the MNIST dataset. It is freely available on the Internet.

Requirements

  1. Python 3
  2. Sklearn
  3. OpenCV 3
  4. numpy
  5. Jupyter-Notebook

Training SVM model

  1. SVM_Classifier.ipynb - This is a ipython notebook so you need jupyter-notebook installed to use this file. Use this file if you want to retrain the model.
  2. digits_cls1.pkl - This is a saved SVM model file.

Digit recognition using OpenCV

dig_rec.ipynb - This is a ipython notebook for recognising handwritten digit in images using OpenCV .This file is using trained SVM model digits_cls1.pkl.

Real time single digit recognition using OpenCV

dig_rec_vid.ipynb - This is a ipython notebook for recognising single handwritten digit using webcam and OpenCV .This file is also using trained SVM model digits_cls1.pkl.

Real time multi digit recognition using OpenCV

multidig_rec_vid.ipynb - This is a ipython notebook for recognising multi handwritten digit using webcam and OpenCV .This file is also using trained SVM model digits_cls1.pkl.

How to use these projects

You can use these projects direct opening the perticular ipython notebook dig_rec.ipynb or dig_rec_vid.ipynb or multidig_rec_vid.ipynb.