Real Time DSU Attendance System Using Facial Recognition
Marking attendance has always been considered a ritual which is followed by almost every institute or organization. Since it has been a trend, it needs to be developed or updated accordingly with the latest technologies. The driving force of this development is the desire to automate, facilitate, sp
2025-06-28 16:34:42 - Adil Khan
Real Time DSU Attendance System Using Facial Recognition
Project Area of Specialization Electrical/Electronic EngineeringProject SummaryMarking attendance has always been considered a ritual which is followed by almost every institute or organization. Since it has been a trend, it needs to be developed or updated accordingly with the latest technologies. The driving force of this development is the desire to automate, facilitate, speed up, save time and efforts; thus, reducing the manual errors and reducing time.
Our system uses facial recognition technology to record the attendance automatically by acquiring images through a high resolution digital camera. The defined algorithm then recognizes faces by comparing the test images with the face images stored in faces/training database. Once the test face matches a stored image, attendance is marked.
The proposed project after implementing on MATLAB and Python will be executed on Raspberry pi.
Project Objectives- The aim of this project is to design an automated, reliable and robust attendance system which reduces manual process errors using face recognition technology.
- The basis of this project lies in using two different algorithms that is Principal Component Analysis (PCA) using Eigen face approach and Convolution Neural Network (CNN) Algorithm based on Triplet Loss Function.
- The usage of both algorithms in recognizing the unknown images will let us know to identify which one is more efficient and why?
The project is implemented using two algorithms; Principle component analysis PCA and convolutional neural network CNN. A high quality web camera is used to capture pictures. The Frame acquisition block of the raspberry pi acquires the pictures from the camera and passes it on as a frame to face recognition pipeline. The Face recognition pipeline involves the algorithms (PCA, ANN) which will be run on Raspberry pi. The Recognition results will be passed to a network stack which contains email or csv file to ensure that a student is marked present or absent for a record. The results are then transferred to Ethernet cable which is connected to DSU network modem (DSU IT structure).
Moreover, for better capturing of Frames, we have initiated the concept of Touch Panel which is connected to HDMI (High-Definition Multimedia Interface) to ensure if students are within the frame.
A printer is also attached, which generates slip of marked attendance for the students.
Benefits of the ProjectReal Time DSU Attendance System provide us with various benefits
- RECORDS ATTENDANCE AUTOMATICALLY
- ANALYZES IF STUDENTS ARE ABSENT
- IT SAVES TIME
- EFFICIENT RECOGNITION UPTO 90%
The attendance is automatically recorded by matching the test image i.e. the image captured through the camera in real time, and is then matched with the training images whose bit file are stored in the raspberry pi. The bit file for the training is generated once using the CNN algorithm, i.e. when images test image is inferred it does not need to perform training again and again thus saving time.
Apart from that, the face recognition algorithm has numerous advantages/applications that includes
- PREVENT RETAIL CRIME
- UNLOCK PHONES FIND
- MISSING PERSONS
- RECORD SCHOOL ATTENDANCE
- FACILITATE SECURE TRANSACTIONS
- Final Deliverable of our project is to automatically mark attendance of our class by using a locally generated DSU Data set implemented on two software; MATLAB and Python, as well as its hardware Implementation on Raspberry Pi 4.
- Attendance of our class will be marked automatically using Raspberry Pi 4, whose Frame acquisition block acquires the pictures from the high-resolution web camera and passes it on as a frame to face recognition pipeline. The Face recognition pipeline involves two algorithms (PCA, CNN) which will be run on Raspberry pi. The Recognition results will be passed to a network stack which contains email or csv file to ensure that a student is marked present or absent for a record. The results are then transferred to Ethernet cable which is connected to DSU network modem (DSU IT infrastructure)
- We have also implemented PCA Algorithm on two software: MATLAB and Python, for AT&T or the ORL Dataset as well as for locally generated DSU Dataset. CNN Algorithm is implemented on Python for locally generated DSU Dataset
- Moreover, for better capturing of Frames, we have initiated the concept of Touch Panel which is connected to HDMI (High-Definition Multimedia Interface) to ensure if students are within the frame.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 45650 | |||
| Raspberry Pi | Equipment | 1 | 16000 | 16000 |
| Web Camera 20 MP | Equipment | 1 | 10000 | 10000 |
| Supply 5V 24A | Equipment | 1 | 500 | 500 |
| Raspberry Pi Case | Equipment | 1 | 500 | 500 |
| Type 'C' Data Cable | Equipment | 1 | 300 | 300 |
| Display Screen (Smart Tab) | Equipment | 1 | 15000 | 15000 |
| Ethernet Cable | Equipment | 1 | 300 | 300 |
| Final Report Printing | Miscellaneous | 150 | 15 | 2250 |
| Interim Report Printing | Miscellaneous | 10 | 10 | 100 |
| Research Papers Printing | Miscellaneous | 50 | 10 | 500 |
| Hard copy Files | Miscellaneous | 2 | 100 | 200 |