OCR PFE Assignment 2023

Introduction

This is a simple project to showcase your skills in the following areas:

• Python or C/C++ programming
• Coding style
• Thinking process
• Solving skills

Problem statement

You are given a set of five images, each of which contains some text in various languages, sizes, orientations, and fonts. Your task is to extract the text from the images and output it in a text file.

• Data is in the input folder.

Image samples

Input 1	Input 3	Input 5

Requirements

• You must use Python or C/C++.
• Evaluate Character Segmentation and Recognition separately.
  • Area of interest detection.
  • Detection per area of interest. (Optional)
• Validation using Character Error Rate (CER) and Word Error Rate (WER).
  • https://github.com/jitsi/jiwer
  • https://github.com/kahne/fastwer
• You are free to use any open-source library or tool.
• Processing speed is a plus.

Evaluation

• The quality of your code will be taken into account.
• The evaluation will be based on the CER and WER of your solution.

Submission

• Fork this repository.

• Add your code to the src folder.

• Generate a output.json file in the output folder.

  • {
      "input_1.png": [
        "This is a sample text",
        "This is another sample text"
      ], # You can add more lines if needed, or more complex data structures
    }

• Add small note in the README.md file on how to run your code.

• When you are done, create a pull request to this repository.

• Request a review from @omarmhaimdat

How to run the code

Download the needed libraries using the following commands:

• pip install paddleocr
• pip install paddlepaddle
• pip install jiwer

Contact

• Omar MHAIMDAT: [email protected]