Smart Attendance System

Face recognition-based attendance system with DeepFace (99%+ accuracy) or LBPH fallback. Captures faces, trains a model, and marks attendance via camera.

Setup

1. Install dependencies:

   pip install -r requirements.txt

2. For higher accuracy (optional): install DeepFace + TensorFlow

   pip install deepface tensorflow

Workflow

1. Capture Faces – Collect face samples for each student
2. Train Model – Train the recognition model on the dataset
3. Take Attendance – Run attendance; press Enter in the camera window to finish

Usage

GUI Launcher (recommended)

python simple_gui.py

• Capture Faces – Add students to the dataset
• Train Model – Train after capturing
• Start Attendance – Take attendance; press Enter to exit and generate the sheet
• Today's Report – View today's attendance report
• Folder buttons – Open dataset, trainer, records, sheets

Command line

# 1. Capture faces (interactive)
python capture_faces.py

# 2. Train the model
python train_model.py

# 3. Run attendance (press Enter to exit)
python attendance.py

Project Structure

• attendance.py – Main attendance system (DeepFace + LBPH)
• capture_faces.py – Face capture (press C to capture, ESC to finish)
• train_model.py – LBPH model training
• semester_register.py – Generates Excel/HTML/CSV register
• simple_gui.py – GUI launcher
• face_alignment_utils.py – Preprocessing (CLAHE, alignment)
• dataset/ – Student face images (Name_RollNo/ folders)
• trainer/ – Trained model (trainer.yml)
• attendance_records/ – Daily attendance CSV/JSON
• attendance_sheets/ – Semester register outputs

Dataset Format

Create folders as Name_RollNo (e.g. John Doe_42). Each folder should have at least 10–30 face images.

Requirements

• Python 3.8+
• Webcam
• OpenCV, pandas, numpy, openpyxl (core)
• DeepFace, TensorFlow (optional, for higher accuracy)