OpenCV OMR Answer Sheet Scanner

A Python application using OpenCV to automatically scan and grade Optical Mark Recognition (OMR) answer sheets. It can process sheets from either a webcam feed or a static image file.

Features

• Scans multiple-choice OMR answer sheets.
• Processes input from a live webcam feed or a pre-existing image file.
• Automatically detects the answer area and an optional score box.
• Uses image processing techniques (Grayscale, Blur, Canny Edge Detection, Contours) to isolate the answer grid.
• Performs Perspective Transform to get a top-down view of the answer grid.
• Detects marked bubbles based on pixel intensity within each answer slot.
• Compares detected answers against a predefined answer key.
• Calculates and displays the final score percentage.
• Overlays visual feedback on the processed image:
  • Highlights detected bubbles (Green for correct, Red for incorrect).
  • Indicates the correct answer bubble if the user's choice was wrong.
  • Draws the calculated score within the detected score box on the original image perspective.
• Provides a stacked view of intermediate processing steps for debugging.
• Allows saving the final graded image with annotations.
• Configurable number of questions and choices per question.

Requirements

• Python 3.x
• OpenCV (opencv-python)
• NumPy

Installation

1. Clone the repository (or download the files):

   git clone https://github.com/donsolo-khalifa/OMRGrading.git
   cd OMRGrading

2. Install the required Python libraries:

   pip install opencv-python numpy

File Structure

• main.py: The main script to run the OMR scanning and grading process. Contains the main loop, image processing pipeline, and grading logic.
• utils.py: A utility module containing helper functions used in main.py for tasks like stacking images, finding/reordering contours, splitting the grid into boxes, and drawing results.
• Scanned/ (Optional/Created on save): Directory where graded images are saved when the 's' key is pressed.

Configuration

Before running, you might need to adjust some settings in main.py:

• webCamFeed: Set to True to use the webcam, False to use an image file.
• pathImage: If webCamFeed is False, set this to the path of your OMR sheet image (e.g., "omr_sheet.jpg").
• cap = cv2.VideoCapture(index): If webCamFeed is True, ensure the index (e.g., 0, 1) corresponds to your desired webcam.
• heightImg, widthImg: Target dimensions for processing the image. Adjust if needed based on sheet aspect ratio or performance.
• questions: The total number of questions on the OMR sheet.
• choices: The number of choices available for each question (e.g., 5 for A, B, C, D, E).
• ans: Crucially, update this list with the correct answer key. It uses 0-based indexing, meaning:
  • 0 = First choice (e.g., 'A')
  • 1 = Second choice (e.g., 'B')
  • 2 = Third choice (e.g., 'C')
  • ...and so on.
  • The length of this list must match the questions variable. Example for 3 questions with answers B, A, C: ans = [1, 0, 2]

Usage

1. Configure main.py as described above (especially webCamFeed, pathImage, questions, choices, and ans).
2. Run the script:

   python main.py

3. Interaction:
   • If using webcam mode, position the OMR sheet in the camera's view.
   • The application will display windows showing the original feed/image, intermediate steps, and the final result with grading overlays and score.
   • Press the 's' key to save the final annotated image to the Scanned/ directory.
   • Press the 'q' key to quit the application.

License

This project is licensed under the MIT License - see the LICENSE file for details.