📸 Face Recognition Login System (Django + DeepFace)

A simple facial recognition system using **Django + DeepFace + OpenCV**.

---

⚙️ Requirements

- Python 3.10 (IMPORTANT ❗)
- Windows 10/11
- Webcam

---

🔍 Check Python Version

py -0

Note: Must be Python 3.10

---

Create Or Switch Virtual Environment

py -3.10  -m venv venv

---

Activate Environment (Windows)

venv\Scripts\activate

---

Install Required Libraries

pip install django deepface numpy

---

pip install opencv-python==4.9.0.80

---

🧠 Install TensorFlow (IMPORTANT for DeepFace)

pip install tensorflow==2.15.0

---

🧠 Install Keras (optional but safe)

pip install keras==2.15.0
pip install tf-keras

---

🚀 Create Django Project

django-admin startproject config .

---

📌 Create App

python manage.py startapp users

---

⚙️ Add App to Django

INSTALLED_APPS = [
    ...
    'users',
]

---

▶️ Run Server

python manage.py runserver

---

📸 Face Recognition Flow

👤 Register Face

• Capture multiple images using webcam
• Save in:

media/faces/username/

---

🔐 Login Face

• Capture image
• Compare using DeepFace
• Return match result

---

🗑 Delete User

media/faces/username/

---

Project Structure

face_login/
│
├── config/
├── users/
├── media/
│   └── faces/
├── venv/
└── manage.py

---

⚠️ Important Notes

✔ MUST use Python 3.10 (3.11/3.13 may break TensorFlow/DeepFace)
✔ Good lighting improves accuracy
✔ Webcam must be accessible
✔ First DeepFace run may be slow (model download)

---

🧠 Common Fixes

❌ TensorFlow error

pip install --upgrade pip

❌ OpenCV camera not opening

Check webcam permissions

❌ DeepFace slow first run

Normal (downloads AI models)

---

👤 1. REGISTER (Save Face)

http://127.0.0.1:8000/users/register/?name=ricardo

---

📸 What happens:

1. Camera opens
2. You press:
    • S = save image
    • Q = quit
3. Images are saved here:

    media/faces/ricardo/
        0.jpg
        1.jpg
        2.jpg

🧠 Result:

User “ricardo” is now registered in the system.

---

🔐 2. LOGIN (Face Recognition)

http://127.0.0.1:8000/users/login/

📸 What happens:

1. Camera opens
2. Takes 1 snapshot
3. Saves temporary image:
    • media/temp.jpg
4. DeepFace compares:
    • temp.jpg vs all saved faces

🧠 Result:

If match found:

{
  "status": "success",
  "user": "ricardo"
}

If no match:

{
  "status": "failed"
}

---

🗑 3. DELETE USER (Remove Face Data)

http://127.0.0.1:8000/users/delete/?name=yourname

🧹 What happens:

Deletes folder:
    media/faces/ricardo/

🧠 Result:

if deleted

{
  "status": "deleted",
  "user": "ricardo"
}

If not found:

{
  "status": "not found"
}

⚙️ SIMPLE FLOW SUMMARY

👤 REGISTER → save face images in folder
🔐 LOGIN → compare camera face with saved faces
🗑 DELETE → remove user folder

---

⚠️ IMPORTANT REMINDERS

✔ Always run server first:

✔ Always activate venv:

venv\Scripts\activate

✔ Use good lighting for face detection

---

🚀 Future Improvements

Real-time video recognition
Face embeddings (faster login)
Database instead of folders
Liveness detection (anti-photo spoofing)
HTML login UI

---

👨‍💻 Author

Name: Ricardo Jr Liera
Project: Face Recognition Login System
Stack: Django + DeepFace + OpenCV
Year: 2026