mycroft-face-wake

Trigger Mycroft without the use of a wake-word.

Simple Demonstration

Short demo: https://www.youtube.com/watch?v=H9P5TRo8nQs&feature=youtu.be

Longer explanation and demo: https://www.youtube.com/watch?v=ytKUTBfjnQI

Description and Motivation

This is a working prototype. It offers no configuration, must be run on the same host as mycroft-core, and all STT is handled by Google's API (via the SpeechRecognition library). It does not replace the wake-word functionality - they can both used at the same time.

I feel that this may offer a more natural way to interact with a voice assistant. Wake words are great, but often when I talk to people, I just make eye-contact and begin talking. This emulates that type of interaction (though again, it does not replace the wake-word functionality).

~~This runs quite nicely on my desktop and laptop, but a Raspberry Pi is not quite powerful enough to handle it effectively (in my brief test, it took roughly 4 seconds to detect a face on the Pi, where it happens virtually instantanesouly on a desktop/laptop). Note that there is also some extra work to get this running on a Pi - it will not work as-is. I may add support at a later date, however, given the performance, it is not a priority.~~

Update: This can work on a pi! See below.

Usage (Tested on Ubuntu 17.10 - see below for Raspbery Pi Support)

1) git clone https://github.com/ChristopherRogers1991/mycroft-face-wake.git 2) cd mycroft-face-wake && pip install -r requirements.txt && pip install . 3) python main.py

Raspberry Pi Support

Demo Video: https://youtu.be/RhQqz_Yy9Fk

This assumes you are running Raspbian on a Raspberry Pi 3.

1) sudo apt-get install python-opencv cmake 2) virtualenv -p python2 --system-site-packages ~/.virtualenvs/mycroft-face-wake 3) source ~/.virtualenvs/mycroft-face-wake/bin/activate 4) pip install dlib 5) git clone https://github.com/ChristopherRogers1991/mycroft-face-wake.git 6) cd mycroft-face-wake && pip install -r requirements.txt && pip install . 7) python main.py

Note that this is using the python-opencv package from the repos, which only seems to work with python2. If you're ambitious, you could probabaly compile it yourself, and get it to work with python3.