Skip to content

Latest commit

 

History

History
74 lines (56 loc) · 3.17 KB

README.md

File metadata and controls

74 lines (56 loc) · 3.17 KB
Raw

facial-recognition

Detects, tracks and tags people with randomly assigned names in a db.

Identified faces in gilette ad

Quick setup

Begin with installing pre-requisites for python packages

sudo apt update && 
DEBIAN_FRONTEND=NON_INTERACTIVE sudo apt install -y \
python3-dev python3-setuptools \
cmake libtiff5-dev libjpeg8-dev libopenjp2-7-dev zlib1g-dev \
libfreetype6-dev liblcms2-dev libwebp-dev tcl8.6-dev tk8.6-dev python3-tk \
libharfbuzz-dev libfribidi-dev libxcb1-dev

Install gStreamer

apt-get install libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev libgstreamer-plugins-bad1.0-dev gstreamer1.0-plugins-base gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-plugins-ugly gstreamer1.0-libav gstreamer1.0-tools gstreamer1.0-x gstreamer1.0-alsa gstreamer1.0-gl gstreamer1.0-gtk3 gstreamer1.0-qt5 gstreamer1.0-pulseaudio

Install python dependencies in a virtualenv

pip3 install virtualenv &&
virtualenv venv &&
source venv/bin/activate &&
pip3 install --upgrade pip && \
pip3 install dlib pymongo numpy tqdm names scikit_image imutils opencv_python Pillow pgi vext vext.gi

Use of CUDA and cudnn is advised.

Development and test system

These scripts were developed and tested on ubuntu 20.04. Due to a lot of dependencies between different dependencies ubuntu 20.04 was chosen as the one offering best compatibility out of the box.

Running

There are a few python scripts, following is an explanation for each

Script Description
streamrecog.py The main scripts which can read from a variety of sources with help of gStreamer.
process-db-manual.py Processes the identified faces and assigns them random names.
chips2png.py Renders a png with the first few hundred faces from the db.

Docker compose

It's recommended the docker compose version to be used as it's most straight forward and comes tested. Pre-requisites are docker, docker-compose and nvidia-docker2. Once installed issue the following commands to build the container and run

su build-docker-images.sh
docker-compose up

By default the container will try to use the web-cam. In case you don't have a webcam then please look in the docker-compose.yaml file for alternative input sources such as a sample youtube video.

The compose file contains 3 containers, which are

  • mongodb listening on port localhost:27017
  • mongo-express, a web ui for mongodb listening on http://localhost:8081
  • this app

Viewing identified faces

Assuming the stream has been running and faces have been accumulated in the database, use the following commands to generate a preview

python3 process-db-manual.py -du mongodb://root:example@localhost:27017/
python3 chips2png.py -du mongodb://root:example@localhost:27017/

Following image shows the mongo express with data gatherd on the gilette ad Mongo Express with some data

Possible applications

This application is a proof of concept that can be extended to use cases of public security by automatically identifying and tagging faces and providing a searchable database.