Optical Character Recognition (OCR) System

This is a simple OCR system that can recognize handwritten digits using an Artificial Neural Network (ANN).

Overview

The system consists of:

• A web-based interface for drawing digits
• A client-side JavaScript component for handling user interactions
• A Python server that processes requests
• A neural network implementation for OCR

Requirements

• Python 3.6+
• NumPy
• A web browser

Installation

1. Install the required Python packages:

pip install numpy

Usage

1. Start the server:

python server.py

2. Open a web browser and navigate to:

http://localhost:5000

3. Draw a digit on the canvas and:
   • Enter the digit value in the text field and click "Train" to train the network
   • Click "Test" to have the network predict the digit
   • Click "Reset" to clear the canvas

How It Works

1. The user draws a digit on the canvas
2. The drawing is converted to a 20x20 pixel representation (400 values)
3. This data is sent to the server
4. The neural network processes the data and returns a prediction

Neural Network Design

The neural network design can be experimented with using:

python neural_network_design.py

This will test different numbers of hidden nodes and report the performance of each configuration.

Architecture

• ocr.html: The web interface
• ocr.js: Client-side JavaScript for handling user interactions
• ocr.css: Styling for the web interface
• server.py: Python server that handles requests
• ocr.py: Neural network implementation
• neural_network_design.py: Script for experimenting with network design