Docker - Installation of Dlib Face Identify How?

Configuration
1

Hi all,

I’m really new to Docker and I just set up my first Docker installation on a Ubuntu 18.04 version with GUI.
I tried to search for installation guides but I don’t get across something that seems to be working for me.

I found this link:
https://hub.docker.com/r/aaftio/face_recognition/ and tried doing a pull which succeded, but from there I’m clueless.
Could someone perhaps guide me through what I need to do on my Docker installation in order to set up the component afterwards in HA?

Much appreciated!

2

You don’t need all that.

I believe it is built into HA.
Just need to add below into your config file

image_processing:
  - platform: dlib_face_detect
    source:
      - entity_id: camera.door
3

Oh, I actually haven’t tried that.
Do you have it running?

Also, I found this one: Dlib Face Identify - Home Assistant which should be the same.

I will try to put it into my configuration file and give it a go :slight_smile:

4

I’ve used the image processing components in past.
They just worked.

EDIT
Dlib Face detect looks for any face
Dlib Face identify looks for specific face
At least that is what I would guess

5

Seems to be accepted in my config. Allthough my current setup is on a Rpi3 and now it runs reeeeeaaaaallly slow.

6

You could try lower resolution or lower frame rate streams.
Still may be slow but it could help

7

I will soon be trying out on a VM with much more power, so I hope that will do the trick.

Thanks for your help!

8

Hey,

I managed to get it working, but have you played around with notifications and how to trigger when a known face is recognized?

9

Yes. but i no longer have these or use this intergration

if you go to “services” under “developer tools” of HA frontend there is likely a service that notifies upon “known face”. It will show you any paramter needed or provided by this. You can utilize this service with the correct data parameter to Trigger your automation.

EDIT

In the Automation Editor you will trigger as an EVENT
EVENT TYPE will be something like image_processing.found_face (this is guess)
DATA PARAMETERS will be like below (also guess since I dont use this and cannot see these in my HA)


{
  "entity_id": "image_processing.dlib_yourcamera",
  "face": [
    "face1",
    "face2",
    "etc....."
  ]
}