rubikproxy 开源 AI 人脸识别考试系统，支持口罩检测与座位分配

Introduction

The Face Recognition Exam System is an AI-powered solution designed to simplify, secure, and streamline attendance and seat allocation during examinations. Leveraging cutting-edge artificial intelligence and facial recognition technologies, this system ensures that only registered and authenticated students participate in exams. It integrates multiple features such as face detection, real-time mask detection, gender recognition, and attendance logging, making it a highly effective and reliable platform.

Features

1. Face Detection and Gender Recognition

• Detects faces from images while also identifying the gender of individuals.
• Uses advanced deep learning facial recognition models for high accuracy.

2. Real-Time Mask Detection

• Identifies whether a face is masked or unmasked.
• Provides feedback to students to remove masks for better identification.

3. AI-Based Student Authentication System

• Verifies identity using pre-enrolled student data via Support Vector Machines (SVM) and face embeddings.
• Prevents unauthorized students from entering exams.

4. Seat Allocation System

• Automatically allocates free seats to authenticated students.
• Optimizes seating plans based on configurable hall layouts.

5. Attendance Logging

• Automatically logs attendance along with seat allocations.
• Stores detailed logs, including date, time, and gender.

6. Enrollment Functionality

• Supports enrolling new students by training the model with their images.
• Maps student images to unique IDs ensuring minimal redundancy.

7. Error Handling and Feedback

• Provides real-time feedback for errors, including non-detection of faces or absence of seat availability.
• Accurately handles various edge cases for smooth operations.

8. Low Resource Dependency

• Lightweight design ensures it can run effectively on machines with minimal resources.

Technologies Used

• Python: Main programming language for implementation.
• OpenCV: Image processing and face detection.
• TensorFlow & Keras: Deep learning model handling and mask detection.
• MTCNN: Multi-task Convolutional Neural Network to detect faces.
• FaceNet: For facial embeddings and authentication.
• DeepFace: For gender recognition and analysis.
• Pandas: For structured data storage and attendance logging.
• Scikit-Learn: Implementing SVM for student identification.

Installation (Linux)

1. Clone the repository:

   git clone https://github.com/rubikproxy/AI-Face-Recognition-Exam-System
   cd AI-Face-Recognition-Exam-System

2. Install the Required Libraries:

   pip install opencv-python numpy scikit-learn keras tensorflow mtcnn keras_facenet deepface
   sudo apt-get update && sudo apt-get install -y libgl1

3. Directory Structure:

   • /work/Models/:
     • Save the pre-trained models like SVM and mask_detector.
   • /work/Attendance/:
     • Stores the CSV file containing attendance records.
   • /work/Student_img/:
     • Contains the student images for enrollment or validation.

Try in Deepnote

How it Works

1. The student's face image or real-time frame is passed to the system.
2. Facial detection is performed to identify and preprocess the face.
3. The system validates whether the student is wearing a mask.
4. The pre-trained SVM model authenticates the student.
5. If authenticated, the system allocates a specific seat from the seating plan.
6. Attendance is logged in a CSV file, recording details such as date, seat number, and gender.

Usage

For Administrators:

• Enrollment: Use the enroll_students() function to add new students to the system.

  student_images = ["/path/to/img1.jpg", "/path/to/img2.jpg"]
  student_ids = ["Student001", "Student002"]
  enroll_students(student_images, student_ids)

For Authentication and Attendance:

```python
face_image = "/path/to/student_current_image.jpg"
result = general_function(face_image)
```
- If a student is recognized, their ID and seat details will be displayed.
- Attendance will be marked automatically in the attendance CSV file.

File Definitions

• svm_model.pkl: Pre-trained Support Vector Machine model for student authentication.
• label_encoder.pkl: Label encoder to map between student IDs and model predictions.
• attendance.csv: Log file for tracking exam attendance.
• mask_detector.model: Pre-trained Keras model for mask detection.

Future Improvements

• Integration with live webcam feeds for real-time attendance management.
• Additional analytics for attendance trends and seat optimization.
• Support for multiple exam halls.
• Improved UI/UX for non-technical users.

Conclusion

The Face Recognition Exam System is a significant step toward automating and improving the efficiency of AI-based attendance and examination management. It not only reduces human intervention but also ensures enhanced security and transparency in the exam process.