ageitgey/face_recognition
Face Recognition
This is a Python facial recognition library designed to detect, encode, and identify human faces in images and video. It functions as a biometric identification tool that converts facial features into numerical encodings to compare and match identities.
The library provides a computer vision command line interface for batch processing face detection and recognition tasks across image directories. It also supports a GPU accelerated vision API that utilizes CUDA and NVIDIA hardware to increase the speed of facial analysis and identification.
Its capabilities cover human face detection and facial landmark mapping for eyes, noses, mouths, and chins. It includes tools for facial identity verification, real-time video recognition, and the training of classifiers to predict the identity of unknown faces.
Pre-configured container images are provided for both CPU and GPU environments to simplify the installation of dependencies.
Features
- Biometric Engines - Functions as a biometric engine that converts facial features into numerical encodings to match identities.
- Face Detection - Locates the pixel coordinates of human faces in photographs using standard or deep-learning models.
- Face Detection - Locates human faces by analyzing gradients of image intensity using Histogram of Oriented Gradients.
- Face Recognition - Matches faces by calculating the geometric distance between face vectors with adjustable thresholds.
- Face Recognition Libraries - Provides a comprehensive Python library for detecting, encoding, and identifying human faces in images and video.
- Facial Landmark Analysis - Detects facial points to rotate and scale images for accurate numerical encoding.
- Facial Landmark Detection - Uses the Dlib toolkit to identify and track key facial landmarks for structural localization.
- Feature Extraction - Transforms facial images into numerical embeddings through convolutional neural network feature extraction.
- Real-Time Facial Recognition - Recognizes known individuals within live video streams by comparing real-time encodings against stored samples.
- Video Input Processing - Processes live webcam feeds and recorded files to identify people with optional bounding box overlays.
- Distance Metrics - Determines identity by calculating the Euclidean distance between face encoding vectors.
- Biometric Face Verification - Matches faces in images or video against known individuals to determine a specific identity.
- Command Line Interfaces - Provides a command line interface for batch processing face detection and recognition tasks across image directories.
- CUDA-Accelerated Vision - Utilizes CUDA and NVIDIA hardware to accelerate heavy matrix computations for faster image analysis.
- Identity Matching - Provides a command-line interface to match faces in unknown images against known identity databases.
- Batch Processing Tools - Provides a command-line tool to locate face coordinates across entire directories of images.
- Face Recognition Training - Provides tools to train classifiers on multiple images per person to predict the identity of unknown faces.
- Batch Identity Matching - Provides a command-line tool to analyze folders of images against known people to output identities.
- KNN Classifiers - Implements identity prediction by comparing facial features against labeled examples using a k-nearest neighbors algorithm.
- GPU Accelerated Computer Vision - Processes multiple images simultaneously using NVIDIA GPUs to accelerate face location discovery.
- GPU Acceleration Configurations - Utilizes NVIDIA GPU hardware and CUDA to accelerate computationally intensive facial recognition tasks.
- Facial Recognition Refinement - Calculates numeric distance between face encodings using custom thresholds to tune match accuracy.
- Biometric Classifiers - Predicts identity of unknown faces by finding the closest matching encodings in a trained dataset.
