Image to Text Conveter in python

Aim

The main aim of the project is to provide a GUI interface for the uploading an image and extracting the text present in the image

Prerequisites

• Tesseract-OCR should be installed on your system

Purpose

For use of Optical Character Recognision in the text extraction from an image

Description

This project is a desktop app which load the text that is included in the image. Libraries used:

• Pillow
• PySimpleGUI
• pytesseract

Workflow of the Project

main() function

The UI is defined under the main function which is called when the window is initialize. The main function contains the:

• Layout for the application
• A loop for checking for events on the window (until the window is terminated)

window.read() function

The window.read() function listens to:

• events that happen on the application window
• values from specific keys

Keys

Keys are given to the ui elements as ids in html. Thus, values in the code is the collection (more like a dictionary for the window) and contains key value pairs as:

values = {
    keys: 'some value'
}

for ex.

values = {
    "-FILE-": "path/to/file"
}

The values are accessed in the similar fashion ( (key, value) pairs where the key is the id of the ui element and value is the innerValue of the element ) in the entire application.

Layout

This is straight forward.

• a 1-D list is row
• a 2-D list is a column

layout = [
    [
        [row1, col1],[row1, col2],
    ]
    [
        [row2, col1],[row2, col2]
    ],
]

ocr_core(filename) function

This function need two things in place to work properly:

• The Tesseract-OCR path should be given properly (see the setup point 4)
• The file that is given should exisit on the computer at the given path

Next is all handled by the pytesseract.image_to_string() function which provided by the pytesseract module

Setup instructions

1. Intitialize a virtual-environment in the directory.
2. now activate the env by
   • ./env/Scripts/activate on windows
   • source env/bin/activate on Linux and Mac
3. Then after, install the dependencies by running pip install requirements.txt
4. Now go and replace your tesseract-OCR path on the line 18 of the image_to_text_converter.py file
5. Now simply run the code by python image_to_text_converter.py

Output

If we have an image such as:

The output will be the extracted text:

You will face many defeats in
life, but never let yourself be
defeated.

MAYA ANGELOU

Conclusion

Thus in this way we have used the pytesseract module and the OCR to detect characters in the given image. This project uses PySimpleGUI which is a simpler yet elegant way for making ui in python.

Author

Nirbhay Chaplot