Register & Capture
Instructors enter student profiles (ID, Name, Major, Gender, Year) and capture real-time webcam photos saved as BLOBs in SQLite.
01 / Introduction
Automating classroom attendance using real-time computer vision.
This project delivers an end-to-end desktop software solution for instructors to streamline student attendance. Using live camera feeds, facial landmarks are extracted and matched against pre-generated embeddings to log attendance in seconds.
The application pairs a modern PyQt6 GUI with lightweight SQLite storage, local Pickle encoding files, and OpenCV face matching to work completely offline without relying on third-party cloud services.
02 / The problem
Manual roll calls waste valuable class time and invite proxy attendance.
Traditional paper-based or manual attendance verification is slow, prone to human error, and allows buddy punching. Instructors need a fast, transparent method to identify students instantly upon entering the classroom.
The challenge was building a desktop system that handles live camera streams smoothly, prevents duplicate logs within a set time window, and manages student records securely.
03 / The solution
A unified desktop suite for registration, encoding, and live verification.
The platform features teacher authentication, a registration portal with webcam image capture, automated vector encoding generation, real-time facial matching, and offline SQLite data persistence.
Vector Encoding
The dlib-backed face_recognition library extracts 128-dimensional facial feature embeddings stored into serialized Pickle reference files.
Live Attendance
Live video streams match faces against stored encodings, incrementing attendance counts while preventing duplicate logs.
SOLUTION / SYSTEM WORKFLOWTEACHER LOGIN → STUDENT REGISTRATION → ENCODING → LIVE MATCHING
04 / Core Engineering & Modules
Robust computer vision integrated with desktop GUI control.
face_recognition & dlib
Deep-learning-based facial feature extraction generates encodings for fast distance-based comparison on live camera frames.
PyQt6 Framework
Multi-window desktop interface featuring teacher login dialogs, student registration forms, and live camera feed containers.
SQLite3 & Pickle Storage
Lightweight, zero-configuration local storage for student demographic data, image BLOBs, timestamp logs, and serialized facial encoding arrays.
05 / Application Pipeline
Execution flow from instructor login to automated attendance.
01LoginInstructor authenticates via the PyQt6 credential dialog.
→
02RegistrationCapture student photo and write profile metadata into SQLite.
→
03EncodingGenerate facial encodings via EncodingGenerator.py.
→
04RecognitionOpenCV matches live webcam frames and updates attendance records.
PIPELINE / SYSTEM WORKFLOWLOGIN → REGISTER → ENCODE → VERIFY
06 / Application Screens
Functional, responsive desktop UI layouts.
Built with PyQt6, cvzone overlays, and custom UI components to display real-time video feeds, student identity overlays, and active class statistics.
07 / Reliability & Constraints
Designed for fast offline processing and consistent verification.
Time Window Thresholds
Prevents repeated attendance increments by tracking last_attendance_time timestamps for each student.
Local Execution
Facial embeddings, captured images, and attendance database records remain on the local filesystem without a required cloud backend.
Modular Codebase
Separated modules for GUI logic, encoding generation, and SQLite handling allow future extensions such as CSV and PDF exports.
08 / Outcomes and Reflections
An efficient approach to automated classroom attendance.
The Face Recognition Attendance System demonstrates how desktop application frameworks and computer vision libraries can be combined to automate attendance without depending on external cloud services.
Key takeaway: Local facial encoding combined with lightweight SQLite storage creates a practical desktop workflow for real-time student identification and attendance logging.
Project details
Built by Septaneuron· Code available on GitHub.
Python 3.11
PyQt6
OpenCV
face_recognition
SQLite3
NumPy / cvzone
View the project on GitHub
Let’s build the next system.
Tell us what you are trying to make clearer, faster, or more capable.
Start a conversation