EliHar-zz / Pattern_recognition

Face recognition using OpenFace

Face Recognition

Install Docker

Installation guide
Docker commands

Pull Docker container

docker pull elihar/openface

Run Docker container

docker run -v [shared/directory/path]:/root/openface/shared -p 9000:9000 -p 8000:8000 -t -i elihar/openface /bin/bash

Go to openface directory

cd /root/openface

Add training data

Add images of people to be classified in directories named after each person. These will be the classes. Then move these directories to ./training-images/

Face detection, cropping, pose detection and alignment

pose_detect_align OR

./util/align-dlib.py ./training-images/ align outerEyesAndNose ./aligned-images/ --size 96

Extract 128 features per face and generate a face representation

gen_rep OR

./batch-represent/main.lua -outDir ./generated-embeddings/ -data ./aligned-images/

Train a face recognition model with desired classifier

train_model [classifier name] OR

./demos/classifier.py train ./generated-embeddings/ --classifier [classifier name]

Classifier name: LinearSvm, GridSearchSvm, GMM, RadialSvm, DecisionTree, GaussianNB, DBN

Recognize an unknown face

recognize [image/path] OR

./demos/classifier.py infer ./generated-embeddings/classifier.pkl [image/path]

Add --multi for recognizing multiple faces in an image

Web Interface

Navigate to /root/web_server

run python server.py

Open 0.0.0.0:8000 in Chrome or Firefox

#####Original resource