Automated Attendance System using ROS2

An intelligent ROS2-based Automated Attendance System that leverages computer vision and face recognition to automate student attendance recording.
This system captures faces in real time, performs anti-spoofing checks, matches them against a pre-stored database, and updates attendance records in an Excel sheet.

---

System Overview

The project is structured into ROS2 nodes that communicate through topics:

• Node1 (Camera & Face Detection)

  • Captures live camera feed
  • Detects faces using Haar Cascade Classifier
  • Performs basic anti-spoofing
  • Publishes detected images

• Node2 (Face Recognition & Attendance Update)

  • Receives detected faces
  • Compares against known student dataset using face_recognition
  • Updates attendance in attendance.xlsx with timestamp, date, name, and match score

• Node3 (Display & Record Keeping)

  • Annotates recognized face image with student details
  • Displays result on screen
  • Saves annotated images in attendance_record/

• Launch File

  • Orchestrates all nodes to run together

---

Features

✔️ Live camera-based face detection
✔️ Anti-spoofing mechanism (basic image variance check)
✔️ Face recognition using deep feature embeddings (face_recognition)
✔️ Automatic attendance logging in Excel (openpyxl)
✔️ Annotated visual feedback with student details
✔️ Modular ROS2 architecture – nodes can run independently or via launch file

---

Algorithms & Techniques Used

• Face Detection → Implemented with Haar Cascade Classifier from OpenCV for locating faces in video frames.
• Face Recognition → Uses the face_recognition library, built on dlib’s deep learning model, which extracts 128-dimensional encodings for each face.
• Face Matching → Recognition is performed by computing the Euclidean distance between encodings. A match score = 1 – distance is derived for easier interpretation.
• Anti-Spoofing Check → A color variance–based method ensures that flat images (e.g., photos shown to the camera) are rejected, reducing spoof attempts.
• Thresholding → Only matches with a score above 0.60 are accepted, balancing accuracy and reliability.
• Attendance Logging → Attendance is updated in an Excel sheet (openpyxl) with timestamp, date, student name, and match score.
• ROS2 Middleware → ROS2 topics provide modular communication between nodes, ensuring scalability (e.g., replacing camera source or adding cloud logging without changing core logic).
• Data Storage →
  • attendance.xlsx → Central attendance record
  • attendance_record → Annotated face images with overlayed names and scores
  • face_detected → Raw cropped face snapshots for debugging

---

Project Structure

automated_attendance_system/
│── launch/
│   └── launch_file.py
│── src/automated_attendance_system/
│   ├── node1.py              # Camera, face detection, anti-spoofing, image publisher
│   ├── node2.py              # Face recognition, attendance logging
│   ├── node3.py              # Display recognized student, save annotated record
│   ├── print_and_launch.py   # Entry script to launch system
│── resource/
│   └── automated_attendance_system
│── setup.py
│── setup.cfg
│── package.xml
│── requirements.txt
│── README.md

---

Installation

1. Prerequisites

• Ubuntu 20.04 / 22.04 with ROS2 Foxy/Humble
• Python ≥ 3.8
• Camera/webcam

2. Install dependencies

sudo apt update
sudo apt install python3-colcon-common-extensions                  ros-${ROS_DISTRO}-cv-bridge                  ros-${ROS_DISTRO}-image-transport

pip install -r requirements.txt

requirements.txt includes:

• opencv-python
• face_recognition
• pandas
• openpyxl
• numpy
• rclpy
• cv_bridge

---

Usage

1. Clone and build the package

git clone https://github.com/<your-username>/automated_attendance_system.git
cd automated_attendance_system
colcon build
source install/setup.bash

2. Run the system

ros2 run automated_attendance_system print_and_launch

3. Stop execution

Press q in the image window to quit.

---

Output

• attendance.xlsx → Logs attendance with

  • Timestamp
  • Date
  • Student Name
  • Match Score

• face_detected/ → Saves detected face snapshots

• attendance_record/ → Stores annotated images with details

---

How It Works (Pipeline)

1. Capture → Camera feed captured via OpenCV.
2. Detection → Haar cascade detects faces.
3. Anti-Spoofing → Color variance check filters spoof attempts.
4. Recognition → Matches face encodings with student dataset.
5. Attendance Marking → Updates attendance.xlsx.
6. Annotation & Storage → Saves/display annotated image.

---

Demo (Conceptual)

• Face Detected: Captured via webcam.
• Recognition: Matches against known dataset.
• Attendance Record: Automatically updated in Excel.
• Annotated Image: Displayed & saved with details.

---

Authors

• Manvendra Pratap Singh

---