EXTRACT

Introduction

EXTRACT is an optical character recognition engine for various operating systems which extracts texts from an image and converts them to plain text.

This model is a very primitive form of the original google tesseract which extracts texts from an image and converts them to plain text.

Modules/Library REQUIREMENTS:

1. os
2. numpy
3. PIL
4. sys
5. keras
6. cropyble
7. cv2
8. shutil

Features

a) Extracts text from input image
b) Works on lowercase, uppercase, numbers and special characters.
c) Saves the output in output.txt to allow search.

How To Run the script:

NOTE1:- The trained model is not provided. So for the very first time run the script as it is. Once the model is trained: COMMENT OUT 'Train_Model()' then run the script for further use.

Run the script on your terminal: 'python3 tesseract.py':

Input Image	Output

Contributors

• Akarsh Malik
• Angad Ripudaman Singh Bajwa

Future Work

1. To add characters of your own, make sure to add them in the train and test dataset
2. Change the output of the softmax layer in Train_Model function to the total number of trained characters.
3. Re-train the model
4. Test your image