Projects/Smart Attendance via Face Recognition
CSE / ITFinal Year

Smart Attendance via Face Recognition

OpenCV + Python attendance system with an admin panel and reports.

Case Study

Problem Statement

Manual roll-call and RFID-card attendance systems are slow, easy to manipulate (proxy attendance), and generate no useful analytics for teachers or managers.

Our Solution

A camera-based system detects and recognizes registered faces in real time, marks attendance automatically with a timestamp, and gives admins a dashboard with daily/monthly attendance reports.

Features

  • Face registration flow (capture 5 angles per person)
  • Real-time recognition using OpenCV + face-recognition library
  • Automatic attendance logging with timestamp and confidence score
  • Admin panel to view, filter, and export attendance reports
  • Duplicate-mark prevention (won't mark the same person twice in a session)

Technology Stack

PythonOpenCVface_recognition (dlib)FlaskSQLiteBootstrap

How It's Built

  1. 1

    Enrollment module stores face encodings per registered user

  2. 2

    Recognition module compares live camera frames against stored encodings

  3. 3

    Attendance service writes a record when a match crosses the confidence threshold

  4. 4

    Admin dashboard reads records and renders filterable reports

FAQ

How accurate is the recognition?

With good lighting and 5 registered angles per person, accuracy is typically 90%+ in the demo dataset. We walk you through tuning the threshold.

Can this work with a laptop webcam?

Yes, no special hardware needed.