OMR Sheet Checker

This project done under PyImageSearch implements an Optical Mark Recognition (OMR) system designed to detect marked cells in a given image of an OMR sheet. It processes the image, detects the contours of the marked cells, and identifies which answers are selected in each question.

Features

• Preprocessing of OMR Sheet: The system processes the input image of the OMR sheet using techniques like thresholding and contour detection to isolate the marked areas.
• Contour Detection: Identifies individual cells in the OMR sheet.
• Answer Recognition: Detects which options (e.g., A, B, C, D) are selected for each question based on the markings.
• Comparison with Correct Answers: The system allows comparison between detected answers and correct answers, making it useful for grading purposes.

Project Workflow

1. Image Preprocessing: The input image is converted to grayscale and processed to highlight the contours using edge detection techniques.
2. Contour Extraction: The contours of each marked region are extracted to identify the individual cells in the OMR sheet.
3. Answer Detection: For each question, the marked option is identified by evaluating which cell has the maximum non-zero pixel count.
4. Result Comparison: The detected answers are compared with the correct answers for evaluation.

Installation

To run this project, install the necessary Python libraries using the following command:

pip install numpy opencv-python matplotlib imutils

How to Use

1. Upload an image of the OMR sheet.
2. Run the code to detect the marked cells.
3. Compare the detected answers with the correct answers.

Code Overview

• OMR Image Preprocessing: Prepares the image by resizing, converting to grayscale, and applying thresholding.
• Grid Detection: Detects contours corresponding to each OMR cell.
• Answer Detection: Uses the non-zero pixel count to determine which option is selected.
• Comparison: Compares detected answers with a predefined answer key.

Example

An OMR sheet is processed as follows:

Original OMR Image	Detected Answers
(Image Input)	(Processed Output)

Demonstration

Original Image v/s Output Image

Actual Selection: [B,E,A,C,B]
Output Selection: [B,E,A,C,B]

Contributing

Contributions are welcome! Fork this repository and submit a pull request with any improvements or bug fixes.

---