📋 2. Dashboard — With Students & Absent List

Once students are enrolled, the dashboard populates with live data. Here 3 students are enrolled and none have given attendance yet for the day — so all 3 appear in the Absent Today panel.

Each absent student is shown as a clickable chip with their name and ID (e.g. MD Naimur Rashid (IRE2101042)). This gives the teacher/admin an instant overview of who hasn't shown up before the class even begins.

📷 3. Face Enrollment — Register New Student

The Face Enrollment page is where new students are registered into the system. The left panel shows a live camera feed with a face detection box drawn around the detected face in real-time.

On the right, you fill in:

• Full Name (e.g. MD Naimur Rashid)
• Student / Person ID (e.g. IRE2101042)
• Department — selected from a dropdown (IRE, CSE, DSE, SWE, etc.)

A progress indicator shows 0 / 30 captured. Once PROCEED ENROLL is clicked, the system starts capturing face samples automatically. The bottom-left panel shows the camera status (CAMERA READY).

⚙️ 4. Enrollment — In Progress (Capturing Samples)

After clicking PROCEED ENROLL, the system begins capturing 30 face images automatically while the student looks at the camera. The golden progress bar fills up as each frame is captured (here: 12 / 30 captured).

The status message below the button reads "Please look at the camera steadily..." — guiding the student to stay still.
Once all 30 samples are captured, InsightFace extracts 512-dimensional embeddings from each image and the SVM classifier retrains automatically in the background — no manual steps needed.

✅ 5. Attendance Verification — First Time Recognition

The Attendance Verification page is where students mark their presence. The camera feed on the left shows the live video with a green circle drawn around the detected face.

On the right, the recognized student's details appear instantly:

• Name: MD Naimur Rashid
• ID: IRE2101042 | Dept: IRE
• Confidence: 85% — how certain the model is about the identity
• Attendance History section shows previous attendance dates and times

The greeting "Welcome back, MD Naimur Rashid! 👋" is shown at the bottom. The student can then click GIVE ATTENDANCE to manually confirm, or use AUTO ON for automatic recording.

🔁 6. Attendance Verification — Already Recorded

If a student tries to give attendance again on the same day, the system detects the duplicate and shows an "Already Recorded" badge (in gold) next to their ID and department.

Here Aamir Khan (DCE0001, Dept: DSE) is recognized at 94% confidence but their attendance was already recorded at 21:00:45. The attendance history panel on the right shows the existing record for 2026-05-22.

This prevents duplicate entries while still showing the student their recognition was successful.

🤖 7. Auto Mode & Manual Mode Buttons

The two action buttons at the bottom of the Attendance page:

• GIVE ATTENDANCE (gold button) — Manual mode. The system recognizes the face and shows the student's details, but waits for a human to click this button to officially record attendance. Useful when the teacher wants to confirm identity before marking.

• AUTO ON (green button) — Auto mode. Once toggled on, the system continuously scans the camera feed and marks attendance automatically the moment it confidently recognizes a face — no button click needed. The button turns green to indicate auto mode is active.

📊 8. Attendance History — Filter & CSV Export

The Attendance History page shows a full log of all recorded attendance entries, sorted newest first. Each row displays: Student ID, Name, Department, Date, Time.

Powerful filtering options at the top:

• 🔍 Search bar — filter by student name or ID in real-time
• 🏢 Department dropdown — show only a specific department
• 📅 Date range (From / To) — narrow down to a specific period
• ✖️ Reset — clear all filters instantly

The EXPORT CSV button (top right) exports the entire filtered view as a .csv file — ready to open in Excel or Google Sheets. The bottom right shows the total record count (3 record(s)).

👥 9. Enrolled Students — Manage Identities

The Enrolled Students page lists every registered student with full details in a table:

• Student ID, Name, Department — identity info
• Enrolled At — exact timestamp when they were registered
• Total Days — how many unique days they've attended (shown in gold)
• Last Seen — the most recent date they gave attendance

A search bar at the top lets you filter by name or ID instantly.
Selecting any row and clicking DELETE SELECTED (red button, top right) removes that student completely — their face dataset, embeddings, and all attendance records are deleted, and the SVM model retrains automatically.

💾 10. Backup & Restore — Data Management

The Backup & Restore page provides two critical data management tools:

📦 BACKUP — Creates a .zip archive containing:

• users.json (enrolled student profiles)
• attendance.json (all attendance records)
• encodings.pkl, recognizer.pkl, labels.pkl (trained model files)

This backup can be used to transfer the entire system to another machine or keep a safe copy before making changes.

⚠️ RESET DATA — Clears all attendance records permanently (with a confirmation dialog). Enrolled students and the trained model are kept intact. This is designed for starting a fresh semester — you keep all enrolled faces but wipe the attendance slate clean.

✨ Features

🗂️ Project Structure

Smart-Face-Attendance/
│
├── 📄 mainT.py                      # Main application — all UI + logic
│
├── 📁 config/
│   └── config.json                  # Paths for model files & camera index
│
├── 📁 data/
│   ├── attendance.json              # All attendance records (per-student history)
│   └── users.json                   # Enrolled student profiles
│
├── 📁 dataset/
│   └── faces/
│       ├── IRE2101042/              # 30 face images per enrolled student
│       ├── DCE0001/
│       └── SWE0001/
│
├── 📁 output/                       # Trained model files (auto-generated)
│   ├── encodings.pkl                # Face embeddings for all enrolled students
│   ├── labels.pkl                   # Label encoder (student ID ↔ integer)
│   └── recognizer.pkl               # Trained SVM classifier
│
├── 📁 logs/
│   └── app.log                      # Error log file
│
├── 📁 screenshots/                  # UI screenshots (used in this README)
│
└── 🤖 volo26n.pt                    # InsightFace model weights

⚙️ How It Works

1. 📸 Enrollment

1. Enter the student's Full Name, ID, and Department
2. The camera