🇺🇸 🇨🇳

Antigen

This is my bachelor graduation design project, a face recognition service based on SeetaFaceEngine. I also provided the public RESTful API. Check this out!

Quick Start

Clone this project via git:

$ git clone https://github.com/Mitscherlich/antigen.git

To run this project locally, you need a pc with ubuntu os installed and at least 2GB memory for OpenCV compiling. If you haven't installed opencv yet, I also provide a script to compile an optimized version.

Then, you need Node.js to run this project. Though apt manager provides a package nodejs for ubuntu but I suggest nvm for installation.

Since npm sometimes generate a wrong dependencies tree, although it is the default package manager came with node.js but yarn is suggested for this project.

Setup MongoDB

This project uses MongoDB to store user data. You can setup a mongo server instance following their official reference or using docker.

To setup a mongo instance with docker, follow this commands below:

# pull `mongo` docker image
$ docker pull mongo
$ docker run -d --name mongo-srv \
# expose mongo default listening port
> -p 27017:27017
# persistence mongo data
> -v ~/db:/data/db
> mongo

If you want to enable user authorization, pass -e MONGO_INITDB_ROOT_USERNAME=<name> and -e MONGO_INITDB_ROOT_PASSWORD=<password> to docker run command and replace mongo with mongo --auth.

Install dependencies

SeetaFaceEngine requires opencv and openblas. Since libopencv-dev not provides the latest version, I suggestion install opencv from it source.

First, you need some system dependencies for compilation:

$ sudo apt-get update
$ sudo apt-get install -fy cmake build-essential dpkg-dev
$ sudo apt-get install -fy libopenblas-dev libprotobuf-dev libssl-dev libgnutls28-dev libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev
# not require if you don't need opencv-python
$ sudo apt-get install -fy python-dev python-numpy libjpeg-dev libpng-dev libtiff-dev libjasper-dev

Then you can use the installation script install-opencv.sh to install an optimized version, or compile a full-function opencv following this commands:

$ mkdir build && cd build
$ cmake -DWITH_QT=ON -DWITH_OPENGL=ON -DFORCE_VTK=ON -DWITH_TBB=ON -DWITH_GDAL=ON -DWITH_XINE=ON -DBUILD_EXAMPLES=ON -DENABLE_PRECOMPILED_HEADERS=OFF ..
$ make -j<jobs> && sudo make install

Jobs usually equals you machine's cpu cores amount times 1.5

Download SeetaFaceEngine models

Downloads models from here or Baidu Disk for mainland users.

Start dev server

You can start this project in development mode by this command:

$ yarn dev

This uses parcel bundler and nodemon for auto restart.

Start prod server

Before you start this project in production mode, you have to build static files:

$ yarn build

Then you can start the server:

$ yarn start

API reference

See docs/API-reference.md for more details.

Common options

This project also accepts some common options by passing environment variables:

• HOST String - The IP address this project will listen. Default is 127.0.0.1.
• PORT String or Number - The port this project will listen. Default is 3000.
• BASE_URL String - The prefix of router and WebSocket, Default is '/'.
• DB_URL String - MongoDB server url. Default is 127.0.0.1.
• DB_PORT String or Number - MongoDB server listeing port. Default is 27017.
• DB_USER String - MongoDB user name. Default is ''.
• DB_PSK String - MongoDB user password. Default is ''.

Docker setup

This project also provides a docker-compose.yml for docker setup. Install docker-compose and execute:

$ docker-compose up -d

Then visit http://localhost:3000 in your browser.

That's it! Have your self :)

License

MIT