https://github.com/thariq-shanavas/Mycroft-FaceRecog

Automating an open source voice assistant to recognise people and give personalised briefing. (Weather, News, etc.)
https://github.com/thariq-shanavas/Mycroft-FaceRecog

Last synced: 3 months ago
JSON representation

Automating an open source voice assistant to recognise people and give personalised briefing. (Weather, News, etc.)

• Host: GitHub
• URL: https://github.com/thariq-shanavas/Mycroft-FaceRecog
• Owner: thariq-shanavas
• License: gpl-3.0
• Created: 2020-01-12T08:34:56.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2020-03-15T12:23:05.000Z (over 4 years ago)
• Last Synced: 2024-04-03T01:44:05.020Z (7 months ago)
• Language: Python
• Size: 189 KB
• Stars: 10
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Mycroft - Face Recognition/Voice Assistant

Automating an open source voice assistant to recognise faces and give personalised briefings. (Weather, News, etc.)

Demo: https://www.youtube.com/watch?v=KHHCviwfhes

The software greets me by the name the first time it sees me in the morning, and gives me weather updates.

Another similar greeting in the evening. The code activates at most two times a day.

# Hardware

* Slightly overclocked Raspberry Pi 3b ($35)

* PS3 Eye Camera ($7)

* Off-the-shelf speaker with a 3.5mm jack.

The easiest way to get started with Mycroft is to use the picroft image - https://mycroft.ai/get-started/

Using a pip installation of OpenCV, you can usually get most features you need without the full installation which takes to the tune of two hours - https://www.pyimagesearch.com/2018/09/19/pip-install-opencv/

For any sort of face recognition, there are usually three distinct phases - 

* Face Detection and Data gathering

* Training the Recogniser

* Face Recognition

MJRoBot does a better job of explaining #1 and #2 than I ever could - https://www.hackster.io/mjrobot/real-time-face-recognition-an-end-to-end-project-a10826

You'll need to run their code to collect facial data and to generate trained models.

You might want to customise the names and indiviual briefings using

`nano face_recognition.py`

The comments should guide you through.

Once you have the trained model to recognise faces, run 

`python face_recognition.py`

Psst. You probably know this already, but if you're running the RPi using ssh, you might want to use 'screen' to open a terminal that keeps running even when the ssh connection is closed - https://linuxize.com/post/how-to-use-linux-screen/

# Acknowledgements

All software is built standing on the shoulders of giants. Thanks to-

* Mycroft developers and community - https://mycroft.ai/about-mycroft/

* MJRoBot's tutorial on face recognition using OpenCV - https://www.hackster.io/mjrobot/real-time-face-recognition-an-end-to-end-project-a10826

* Adrian Rosebrock's tutorial on OpenCV installation using pip - https://www.pyimagesearch.com/2018/09/19/pip-install-opencv/