Face Recognition based on FaceNet paper by Google researchers

A simple Face recognition AI coupled with Node.js server.
Following Technologies were used while building the software

Machine Learning

• Python
• Tensorflow
• Keras
• Anaconda

Image Processing

• dlib
• OpenCV

Server

• Node.js
• Express.js

Pub/Sub

• Redis

API end point

Post localhost:3000/

• Request Body to train

  You can send upto 10 images for each person
  image is an array of image files

note: you have to prepend “images/“ to each file name
I am planning to change this in future

{
    "type": "train",
    "data": {
        "name": "String",
        "image": ["images/file1", "images/file2"]
    }
}

• Request Body to test

note: you have to prepend “images/“ to each file name
I am planning to change this in future

{
    "type": "test",
    "data": "images/file"
}

How to run on your own machine

1. Before You start make sure you have every thing setuped

   • Install Anaconda
   • Create a Conda environment
   • Install Keras and Tensorflow
   • Install Opencv and dlib
   • Install Node.js
   • Install Redis

2. Next cd to faceReco dir

   cd /path/to/faceReco

3. Start redis server

   redis-server --daemonize yes

4. Start Node server

   npm install

   node server.js

5. Activate Conda environment

   source activate environmentName

6. Start AI

   python pubsub.py