😏Face Recognition Tool😀

via GIPHY

Project Overview

This Python-based project is a face recognition system that leverages MTCNN for face detection and FaceNet for generating face embeddings. Recognized faces are classified using an SVM classifier. The system offers real-time face detection, extraction of face embeddings, and recognition via webcam.

Features

• Face Detection: Detects faces in real-time using MTCNN.
• Face Recognition: Recognizes faces through FaceNet embeddings and an SVM classifier.
• Real-time Recognition: Utilizes webcam for live face detection and classification.
• Dataset Support: Handles a dataset of labeled faces for training.

Project Structure

.
├── dataset/        # Folder containing labeled images of faces
├── facenet_model/   # Pre-trained FaceNet model
├── requirements.txt # List of required dependencies
└── face_recognition.py # Main Python script for face recognition

How It Works

1. Face Detection: The MTCNN detector identifies faces within an image frame.
2. Face Embeddings: Each detected face gets resized and processed through FaceNet to generate a 128-dimensional embedding vector.
3. Classification: Embeddings are classified using a linear SVM classifier.
4. Real-time Recognition: The system recognizes faces in real-time, labeling them within the webcam feed.

Installation

1. Clone the repository:

git clone [email protected]:Ruthwik2610/Face_recognition_using_FaceNet_and_MTCNN.git

cd Face_recognition_using_FaceNet_and_MTCNN  

Install dependencies:

pip install -r requirements.txt 

Setting up the dataset:

Create a folder named dataset/. Place images in subfolders within dataset/, where each subfolder is named after the person in the images (e.g., dataset/Alice/, dataset/Bob/). Running the Script

python face_recognition.py

Dataset Format

The dataset should consist of images of faces organized in subfolders. Each subfolder represents a unique person. Ensure images are clear and frontal-facing. example:

dataset/
├── Alice/
│   ├── alice_1.jpg
│   ├── alice_2.jpg
├── Bob/
│   ├── bob_1.jpg
│   ├── bob_2.jpg

Code Overview

Face Loading Class: Loads face images from the dataset, detects faces, and extracts face embeddings.
SVM Classifier: Trains an SVM classifier on the generated face embeddings.
Real-time Recognition: Uses OpenCV to capture webcam input and recognize faces in real-time.