[SOLVED] Successful face detection with demographics test?

Has anyone had luck with a face detection + demographics app? I get the following error trying to use the detection engine with type: demographics.

Face <Buffer ...>

When the app starts I get a message that says No Detection Component Available for Ping. Does that mean that demographics are not yet available on matrix-os?


I solved this by upgrading malos and then running malos_eye

Thanks @kandizzy for updating this…

We’re updating daily guys, if you guys have any issues, just make sure all the packages are upgraded in general as well.


// malos, malos_eye, hal
sudo apt-get update && sudo apt-get upgrade


If you’re using OS…

git pull
npm install
git submodule update --init


The face detection test is still not working for me after doing the suggested steps above (I used the face-demo app, described here: https://www.youtube.com/watch?v=8jkWwum8O8Q)


  • installed the latest matrix-os
  • re-ran sudo apt-get update and sudo apt-get upgrade to make sure that all Debian packages are updated
  • started malos_eye

When starting the face-demo app, I get the following logs from malos_eye process:

MALOS eye starting

You can query specific driver info using port 22012.
Registered driver GESTURE with port 22013.
Configured camera id: 0
Configured resolution 640x480
Gesture initialization done
New delay between updates for GESTURE is 1000 ms.
Configuring detections
Activating FACE
VIDIOC_STREAMON: Invalid argument
VIDIOC_STREAMON: Invalid argument
VIDIOC_STREAMON: Invalid argument
VIDIOC_STREAMON: Invalid argument

Any idea what the problem could be? I am using Matrix CLI v1.2.7 and Raspberry Pi 3 with PiNoir v2 camera module.

Thank you!

Hi again!

After I wrote this post, I decided to also restart the Raspberry Pi and retry. It worked after that! :slight_smile:

Can either of you explain what you launched to see the video stream with the overlay? I believe the app is running and from the stdout where node.js is running I can see it echo [ ‘detection’ ]
but from the example videos it looks like you are watching the detection in realtime somehow. Thanks for any help on this!