https://github.com/imdeepmind/digitrecognition

Simple Digit Recognition Web App using CNN
https://github.com/imdeepmind/digitrecognition

artificial-intelligence artificial-neural-networks classification computer-vision convolutional-neural-networks deep-learning flask keras machine-learning mnist python3

Last synced: 28 days ago
JSON representation

Simple Digit Recognition Web App using CNN

• Host: GitHub
• URL: https://github.com/imdeepmind/digitrecognition
• Owner: imdeepmind
• Created: 2018-11-03T12:57:58.000Z (about 6 years ago)
• Default Branch: master
• Last Pushed: 2022-11-22T03:05:40.000Z (almost 2 years ago)
• Last Synced: 2023-03-05T10:34:05.580Z (over 1 year ago)
• Topics: artificial-intelligence, artificial-neural-networks, classification, computer-vision, convolutional-neural-networks, deep-learning, flask, keras, machine-learning, mnist, python3
• Language: Python
• Homepage: https://imdeepmind.com/DigitRecognition/
• Size: 32.7 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 8
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Digit Recognition

Digit Recognition is a simple project that predicts handwritten digits.

![Screenshot from 2019-08-04 20-57-30](https://user-images.githubusercontent.com/34741145/62425763-84f15b00-b6fa-11e9-8c6e-03fa9343a86b.png)

## Branches

In this repo, there are 3 branches

1. master - This branch contains the main web app. I've used Flask for this part of the project.

2. cnn-keras - This branch contains the main CNN for making the classifier

3. gh-pages - This branch contains a simple Web app created using Bootstrap 

## Programming Languages and Libraries used

Python, Flask, Keras, Numpy, Pandas, HTML, CSS, JavaScript, jQuery, and Bootstrap

## gh-pages Branch

This branch contains a very simple web app. In this app you can draw a simple digit (0-9) and then click upload. It will recognize the digit. 

To run part of the project locally just run the index.html file. 

## cnn-keras Branch

This branch contains the python code for training a complex CNN. To run this part of the project locally, first install tensorflow and keras, and then extract the mnist_png.tar.gz file. Finally run the cnn.py file.

Link to colab notebook - https://colab.research.google.com/drive/1JOLiEy8uc-8Jr1i2ftbjigjsqFbfYdtQ

## master Branch

This branch contains a very simple REST API developed in Flask. This API just uses the model and predict the result.

## Demo

https://imdeepmind.com/DigitRecognition/