Face Detection & Verification - FastAPI

This repository contains a Face Detection & Verification system built using FastAPI. The model utilizes One-Shot Learning with a deep learning-based FaceNet model to identify and verify individuals.

Features

• Face detection using OpenCV
• Face embedding extraction using FaceNet
• One-shot learning for face verification
• FastAPI-based REST API for easy interaction

---

Installation

1. Clone the Repository

git clone https://github.com/Moaz0009/Face-Detection-Verification-FASTAPI.git
cd Face-Detection-Verification-FASTAPI

2. Install Dependencies

Make sure you have Python installed, then run:

pip install -r requirements.txt

3. Run the FastAPI Server

uvicorn main:app --reload

The API will be available at: http://127.0.0.1:8000

---

API Endpoints

1. Health Check

Endpoint: GET /

Response:

{"health_check": "OK"}

2. Get API Information

Endpoint: GET /info

Response:

{"name": "Face_Detect_Verify", "description": "A model using One-Shot Learning to Identify and Verify Individuals"}

3. Upload a Face Image

Endpoint: POST /upload_image

Request:

• name (form data) - Name of the person
• image (file) - Image file to upload

Response:

{"status": "success", "message": "John Doe added to database"}

4. Detect & Verify Face

Endpoint: POST /detect_verify

Request:

• image (file) - Image file to verify

Response:

{"status": "success", "faces": {"John Doe": [x, y, w, h]}}

---

How It Works

1. Face Embeddings Generation:
   • When an image is uploaded, the model extracts face embeddings using FaceNet and stores them in a database.
2. Face Verification:
   • When an image is sent for verification, the model extracts embeddings and compares them with stored embeddings using Euclidean distance.
3. Face Detection:
   • Uses OpenCV’s Haar cascade classifier to detect faces before encoding.

---

Dependencies

• FastAPI
• OpenCV
• TensorFlow/Keras
• NumPy

Install them using:

pip install fastapi opencv-python-headless tensorflow numpy uvicorn

---

Future Improvements

• Improve face recognition accuracy using different distance metrics.
• Add support for multiple faces in a single image.
• Implement database persistence instead of in-memory storage.

---

Author

Moaz Abdeljalil

Feel free to contribute or report issues! 🚀