Upload a bunch of photos of you, your friends, your family or your favorite celebrity. Train the system, and recognize their other photos anytime, anywhere.

The environment set up process is designed keeping specific underlying operating system in mind. Will suggest to follow the setup process below on an Ubuntu machine preferably Ubuntu 14.04 or Ubuntu 16.04 on which the below process is tested and working OK.

• Making sure that the base OS has all the latest packages.

sudo apt-get update
sudo apt-get upgrade

• Installing developer tools for linux system.

sudo apt-get install build-essential cmake git pkg-config

• Installing supporting libraries for OpenCV

  Installing support to load various types of Images

  sudo apt-get install libjpeg8-dev libtiff4-dev libjasper-dev libpng12-dev
  

  Installing support for displaying images on screenshot

  sudo apt-get install libgtk2.0-dev
  

  Installing support for image and video processing

  sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev libv4l-dev
  

  Installing image processing routines optimization libraries

  sudo apt-get install libatlas-base-dev gfortran
  

• Installing pip, a popular python package manager

wget https://bootstrap.pypa.io/get-pip.py
sudo python get-pip.py

• To avoid possible clashes with any existing development environment install virtualenv and virtualenvwrapper

sudo pip install virtualenv virtualenvwrapper
sudo rm -rf ~/.cache/pip

• Adding path variables to .bashrc for setting up virtualenv and virtualenvwrapper

  Add the following lines to ~/.bashrc file using simple text editors of choice

  # virtualenv and virtualenvwrapper
  export WORKON_HOME=$HOME/.virtualenvs
  source /usr/local/bin/virtualenvwrapper.sh
  

  Execute below command to reset the ~/.bashrc file and take effect into bash terminal

  source ~/.bashrc
  

  Execute below command to create a new virtual environment

  mkvirtualenv opencv
  

• Setting up the virtual environment for deployment

  Installing python development tools

  sudo apt-get install python2.7-dev
  

  Installing numpy for processing images as a multi-dimensional arrays.

  sudo pip install numpy
  

  Downloading OpenCV Supporting Modules Source

  cd ~
  git clone https://github.com/Itseez/opencv.git
  cd opencv
  git checkout 3.0.0
  

  Downloading OpenCV Supporting Modules Source

  cd ~
  git clone https://github.com/Itseez/opencv_contrib.git
  cd opencv_contrib
  git checkout 3.0.0
  

  Setting up the build for compilation

  cd ~/opencv
  mkdir build
  cd build
  cmake -D CMAKE_BUILD_TYPE=RELEASE \ -D CMAKE_INSTALL_PREFIX=/usr/local \ -D INSTALL_C_EXAMPLES=ON \ -D INSTALL_PYTHON_EXAMPLES=ON \ -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib/modules \ -D BUILD_EXAMPLES=ON ..
  

  Compiling the code of OpenCV and Supporting Modules

  make -j5
  

  Installing and setting up OpenCV and the Modules

  sudo make install
  sudo ldconfig
  

  Configuring the virtual environment with installed OpenCV packages

  cd ~/.virtualenvs/cv/lib/python2.7/site-packages/
  ln -s /usr/local/lib/python2.7/site-packages/cv2.so cv2.so
  

  Verifying successful installation of the environment

  workon opencv
  python
  >>> import cv2
  >>> cv2.__version__
  '3.0.0'
  

• Installing Flask for Web Application Deployment

sudo pip install flask

• Installing Pilliw for Image saving and loading

sudo pip install Pillow

• Installing JSON parsing

sudo pip install simplejson

• Installing MySQL for Information Storage

sudo apt-get install mysql-server

• Setting up the database

  Login as root user

  mysql -u root -p
  <root account password that you set while installing mysql>
  

  Create a new database

  create database LabelInfo
  use LabelInfo
  

  Create table for the label information

  CREATE TABLE `LabelInfo`.`tbl_details` (`label_id` BIGINT NOT NULL AUTO_INCREMENT, `first_name` VARCHAR(45) NULL, `last_name` VARCHAR(45) NULL, `location` VARCHAR(45) NULL, `birthdate` DATE NULL, PRIMARY KEY (`label_id`));
  

• Giving our code the information about mysql connectivity

  Open Terminal and navigate to Home directory

  cd ~
  

  Download the project from github repository

  git clone https://github.com/keyurgolani/OnYourFace.git
  

  Find app.py into the project

  cd ~/OnYourFace/OnYourFace/
  

  Open app.py in any terminal example gedit

  gedit ~/OnYourFace/OnYourFace/app.py
  

  Change the mysql database user information to your instance information on lines 15 through 18

  app.config['MYSQL_DATABASE_USER'] = 'root'
  app.config['MYSQL_DATABASE_PASSWORD'] = '<The password you set for root user while installing mysql>'
  app.config['MYSQL_DATABASE_DB'] = 'LabelInfo'
  app.config['MYSQL_DATABASE_HOST'] = 'localhost'
  

• Starting the Application

Making sure that we are on right virtual environment

workon opencv

Starting the Python Flask Application Server

cd ~/OnYourFace/
python ~/OnYourFace/runserver.py

• Accessing the Application

  • The application, with the default configurations, should be up and running on localhost at 2016 port.
  • Open your browser and access the application with following URL in address bar.

  https://localhost:2016/
  

• Training the system

  • Go to Upload tab from the header of the application.
  • Fill in the information for the person you want to upload photos of.
  • Select multiple photos of the person in which the person's frontal face is visible. The more photos the better. But upload the right photos.
  • Click the upload button.
  • Repeat this step for each separate individual you want to train the application for.
  • At the end, go to the Train tab on the header, click Train, sit back and relax for a while.

• Identifying a person

  • After training the application with several photos, you will automatically be redirected to the recognize page. If not, you can access it through Recognize tab on the header.
  • Select the photo to be recognized and click Recognize button at the bottom and let the application work its magic.

Have a little fun with the application. Suggest more functionalities. I will try to integrate more and more as suggestions come in. Contributions are always welcome. Let's make it big.

If you try to recognize a person that the application has not been trained, it will try to find the closest feature match from the people it has been trained with.

• Integrate video capabilities for training and recognizing the face.
• Integrating object recognition feature.
• Integrating OCR facility.
• Integrating nudity recognition.
• and suggested features of course.