https://github.com/saadarazzaq/autograde

Automated MCQ Grading with Computer Vision and Optical Mark Recognition (OMR) Technology✅
https://github.com/saadarazzaq/autograde

automation computer-vision image-processing optical-mark-recognition

Last synced: over 1 year ago
JSON representation

Automated MCQ Grading with Computer Vision and Optical Mark Recognition (OMR) Technology✅

• Host: GitHub
• URL: https://github.com/saadarazzaq/autograde
• Owner: SaadARazzaq
• Created: 2023-10-25T17:57:49.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2024-12-23T04:09:25.000Z (over 1 year ago)
• Last Synced: 2025-01-23T16:14:27.585Z (over 1 year ago)
• Topics: automation, computer-vision, image-processing, optical-mark-recognition
• Language: Python
• Homepage:
• Size: 242 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# AutoGrade  

**Automated MCQ Grading with Computer Vision and Optical Mark Recognition (OMR) Technology** ✅  





### Overview  

AutoGrade is a smart solution to grade multiple-choice questions (MCQs) automatically. It uses computer vision and Optical Mark Recognition (OMR) to make grading faster and more accurate, saving time and reducing errors.

### Key Features  

- **Smart Image Analysis**: Detects and checks marked answers with image processing.  

- **Precise Recognition**: Finds and reads filled bubbles or checkboxes accurately.  

- **Handles Many Sheets**: Grades multiple answer sheets at once.  

- **Flexible Designs**: Works with different types of answer sheet layouts.  

- **Quick Results**: Gives instant scores and useful feedback.

### Technology Stack  

- **Programming Language**: Python  

- **Libraries**:  

  - OpenCV for processing images  

  - NumPy for calculations  

  - Pandas for organizing data  

- **Backend**: FastAPI for APIs and app setup  

- **Testing**: PyTest for checking code quality  

### Challenges Solved  

- Reduces mistakes from manual grading  

- Speeds up the grading process  

- Handles large numbers of answer sheets efficiently  

### Outcome  

AutoGrade makes grading easy and accurate, helping teachers and examiners save time and focus on other tasks.

---

## Approach  

### 1. Image Preprocessing  

- Load and resize the input image for consistent analysis.  

- Convert to grayscale to simplify the image.  

- Apply blur to reduce noise.  

- Detect edges with the Canny edge detector.  

### 2. Finding Rectangles  

- Find all contours (shapes) in the image.  

- Select rectangle shapes based on size and structure.  

- Sort the rectangles by size and pick the important ones.  

### 3. Fixing the View  

- Get the corners of each rectangle.  

- Use perspective transformation to view the rectangle flat like a paper.  

- Crop to keep only the part with answers.  

### 4. Recognizing Marks  

- Change each cropped rectangle to grayscale.  

- Use binary thresholding to highlight the marks.  

- Check the density of pixels to see which answers are filled.  

### 5. Showing Results  

- Use a Tkinter window to show all steps and results in one place.  

### 6. Tools and Functions  

- Functions to find rectangles and order their corners.  

- Tools to filter and display images step by step.  

---

### Contact  

For more details, feel free to reach out via email or connect on LinkedIn.