Skip to content

This project builds a smart face recognition system using Python, dlib, and pre-trained models. It detects faces with HOG or CNN methods, maps facial landmarks, and compares face encodings for identification. Ideal for security, photo management, and video surveillance applications.

Notifications You must be signed in to change notification settings

maulikthakur/Facial-Recognition-System

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Facial-Recognition-System

This project implements a face recognition system using the dlib library and pre-trained models from the face_recognition_models package. The system is capable of detecting faces in images, extracting facial landmarks, and computing facial encodings to identify or compare faces based on their similarity.

Key Features:

Face Detection:

The project supports two detection models: HOG-based detector (faster but less accurate) CNN-based detector (more accurate, especially for smaller faces, but requires GPU acceleration for optimal performance). Faces are detected and represented as bounding boxes in the image. Facial Landmark Detection:

Supports both 68-point and 5-point facial landmark predictors to capture the geometry of the face for alignment and recognition. Face Recognition:

Uses a pre-trained model to generate face encodings, which are high-dimensional feature vectors unique to each face. These encodings can be compared to calculate similarity between faces using Euclidean distance, enabling the system to identify or verify individuals. Image Handling:

Loads image files using the Pillow library and converts them into NumPy arrays for processing. Ensures that face detection bounding boxes stay within the bounds of the image. Distance Calculation:

Compares multiple face encodings to a known face encoding and returns the similarity scores, which are used to determine how closely faces match. Technologies Used:

Python: Core language for implementation. dlib: Library for face detection, shape prediction, and face encoding. Pillow (PIL): Used for loading and manipulating image files. NumPy: Handles numerical operations and array manipulation. face_recognition_models: Pre-trained models for face detection and recognition. Applications:

Identity verification for security systems. Facial recognition in photo management applications. Real-time face recognition for video surveillance. This project provides a robust, scalable solution for face recognition tasks by leveraging both classical (HOG) and deep learning (CNN) approaches, making it adaptable for different performance needs.

About

This project builds a smart face recognition system using Python, dlib, and pre-trained models. It detects faces with HOG or CNN methods, maps facial landmarks, and compares face encodings for identification. Ideal for security, photo management, and video surveillance applications.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages