
Smart Attendance Tracking System
Automating Workplace Attendance with Real-Time Face Detection
“Attendance should be automatic, secure, and traceable from detection to analytics.”
System Overview
This project is a desktop smart attendance application that performs live face recognition from webcam input, validates liveness before marking attendance, and stores records in a local SQLite database. The core runtime is built around a Tkinter interface with tabs for dashboard, reports, face management, analytics, settings, and help.
Recognition and Anti-Spoofing
The attendance loop supports a standard face_recognition path with optional deep-model recognition and YOLOv8-assisted detection. Before attendance is written, the system runs anti-spoofing checks using Eye Aspect Ratio blink detection and landmark movement scoring, reducing false acceptance from static photos or non-live input.
Data, Reports, and Analytics
Attendance persistence uses deduplicated day-level inserts in SQLite, so each person is marked once per day. Reporting and analytics are handled with pandas, matplotlib, and scikit-learn metrics (including ROC and AUC flows), with CSV export support and generated graphs for review and benchmarking.
End-to-End Pipeline
On startup, the app initializes configuration, optional models, dlib landmarks, logging, and database tables. Enrollment captures and validates face samples, stores them in known_faces, and reloads encodings. During live attendance, frames are processed for detection, recognition, liveness, threshold checks, and attendance marking, then surfaced in dashboards and analytics views.
Project Gallery
10 items




Explore More Projects
Swipe or click arrows to see other work
















