Skip to content

e-Employee Attendees System with face identification using Facebook DeepFace model

License

Notifications You must be signed in to change notification settings

danieldanuega/halo

Repository files navigation

Welcome to halo project

Face identification system built on top of Facebook DeepFace model and openCV.

Dependencies

  • Offline = -
  • Semi-Online = halo-services

About

This project is an employee attendees system to identify employees face when they come to office. This repo is a standalone system for offline and semi-online identification solution. Also this repo contains the model creation by using tensorflow 2 and keras model. This project inspired by serengil/deepface project.

2 Solutions

This repo has 2 solutions:

  • halo.py / Offline: system will run the model on the device and also compute the similarity on the device.
  • halo_serving.py / Semi-Online: system will request for image embedding to tensorflow-serving (halo-services) server and then compute the similarity on the device.

Usage

  1. Install the requirements using pip install -r requirements.txt.
  2. Add database folder on root project.
  3. The default solution for this project is Offline solution. If you want to use Semi-Online solution, please change the import statement in rekog.py from from halo import FaceRecognition --> from halo_serving import FaceRecognition.
  4. Run the recognition script python rekog.py to start live verification based on the faces in database folder.

About

e-Employee Attendees System with face identification using Facebook DeepFace model

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages