📸 SnapClass

Making attendance faster using AI

Face recognition and voice biometrics that turn a single class photo (or a quick roll-call) into an attendance record.

Live App · Landing Page · Frontend Repo

---

📖 Overview

Taking attendance by hand eats into class time and is easy to get wrong or fake. SnapClass is an AI-powered attendance system where a teacher snaps one or more photos of the classroom, and the app identifies every enrolled student by face. For a different flow, students can say "Present" one by one and the app matches each voice against stored voice embeddings.

Teachers create a subject, share a QR code or join link, and students enroll themselves, registering their face (and optionally their voice) once. From there, attendance is a few clicks, with a review step before anything is saved.

✨ Features

For teachers

• 🔐 Secure registration and login (passwords hashed with bcrypt)
• 📧 Account recovery by email: forgot your password? Enter your email and get a reset link in your inbox to set a new one. Forgot your username? Enter your email and get it sent to you
• 📚 Create and manage subjects, each with a unique subject code
• 📱 Share a join link or QR code so students enroll in seconds
• 📸 FaceID attendance: upload one or more class photos and run face analysis
• 🎙️ Voice attendance: students speak in sequence and are matched by voice embedding
• ✅ Review the detected attendance report, then confirm and save (or discard)
• 🗂️ Attendance records per subject, with confidence scores and CSV export

For students

• 🔗 One-click enrollment through a QR code or join link (auto-enroll after login)
• 🙂 Login with password or FaceID
• 🎤 Optional voice enrollment during registration
• 📊 Personal dashboard showing enrolled subjects and attendance

🎬 How It Works

Teacher creates subject ──► Shares QR / join link ──► Students enroll + register face (and voice)
                                                                    │
Teacher takes class photo(s) or runs voice roll-call ◄──────────────┘
        │
        ▼
Face / voice matched against stored embeddings
        │
        ▼
Teacher reviews report ──► Confirm & save ──► Records stored in Supabase

🔑 Account Recovery Flow

Teacher clicks "Forgot password" ──► Enters registered email ──► Reset link sent to inbox
        │                                                              │
        └──────────────► Opens link ──► Sets a new password ◄──────────┘

Teacher clicks "Forgot username" ──► Enters registered email ──► Username sent to inbox

🧰 Tech Stack

Layer	Tools
App framework	Streamlit
Face recognition	face_recognition, dlib
Voice biometrics	Resemblyzer, librosa
Database & auth storage	Supabase (PostgreSQL)
Security	bcrypt
Enrollment QR codes	segno
Data & ML utilities	numpy, pandas, scikit-learn, Pillow
Landing page	Flask, deployed on Vercel (separate repo)

📁 Project Structure

snapclass/
├── app.py               # Entry point: routing between home, teacher and student screens
├── requirements.txt     # Python dependencies
└── src/
    ├── components/      # Reusable UI pieces (e.g. auto-enroll dialog)
    ├── database/        # Supabase connection and data access
    ├── pipelines/       # Face recognition and voice processing pipelines
    ├── screens/         # Home, teacher and student screens
    ├── services/        # Business logic (auth, account recovery, etc.)
    └── ui/              # Shared styling and UI helpers

🚀 Getting Started

Prerequisites

• Python 3.10+
• A Supabase project
• A webcam and microphone (for FaceID login and voice enrollment)

Installation

# 1. Clone the repository
git clone https://github.com/Ariesjeev/snapclass.git
cd snapclass

# 2. Create and activate a virtual environment
python -m venv venv
source venv/bin/activate        # Windows: .venv\Scripts\activate

# 3. Install dependencies
pip install -r requirements.txt

Configuration

Add your Supabase credentials in .streamlit/secrets.toml:

SUPABASE_URL = "your-project-url"
SUPABASE_KEY = "your-anon-or-service-key"
# Public URL where the running Streamlit app can receive the recovery link.
APP_URL = "http://localhost:8501"

# Gmail SMTP: use a Google App Password, not your normal Gmail password.
SMTP_HOST = "smtp.gmail.com"
SMTP_PORT = 587
SMTP_USERNAME = "your-gmail-address@gmail.com"
SMTP_PASSWORD = "your-16-character-app-password"
SMTP_FROM = "SnapClass <your-gmail-address@gmail.com>"

Make sure your Supabase tables match what the app expects (teachers, students, subjects, enrollments and attendance logs).

Run

streamlit run app.py

The app opens at http://localhost:8501.

🖥️ Usage

1. Teacher: register, log in, and create a subject from Manage Subjects.
2. Click Share Class Link and send the QR code or link to students.
3. Students: open the link, register (face plus optional voice), and enroll.
4. Teacher: open Take Attendance, pick a subject, add class photos and run Face Analysis, or use Voice Attendance.
5. Review the attendance report and Confirm & Save.
6. Check history any time under Attendance Records.

Forgot your login details? (teachers)

• Forgot password: click Forgot password, enter your registered email, open the reset link sent to your inbox, and set a new password.
• Forgot username: click Forgot username, enter your registered email, and your username is sent to you.

🗺️ Roadmap

• Liveness detection to block photo spoofing
• Per-student attendance analytics and low-attendance alerts
• Bulk CSV export across subjects
• Docker setup for easier local install

🙏 Acknowledgements

This project was built while following the SnapClass tutorial by Apna College, then extended and deployed on my own. Face recognition is powered by face_recognition and voice embeddings by Resemblyzer.

👤 Author

Jeevan Bikash Sahoo: Full Stack Developer & AI Engineer GitHub: @Ariesjeev

---

If you found this useful, consider giving it a ⭐