Hey there!
I wonder if this scenario is not too hard to configure:
Camera inside main door captures me, AI platform identifies me, the door opens, alarm system turns off and some other stuff.
Is it possible to use something like Microsoft Face Identify or Dlib Face Identify or something harder like tensorflow on other raspbery pi with your own neural network.
This is so exiting, but I do not have enough knowledge to work with tensorflow (i am not a programmer), would like to get some advices.
Also would like to know how accurate are those platforms.
I will setup the same thing using our Xeoma camera server. It has a very skilled face detection module that can be trained on faces. When a face is recognized, it sends a message via http request to node red. Node red can then be setup with an automation that do the things you want. Is this what you want? I have had no time to set it up yet but I know it works
Nope, not only pi. I thought it would be ok for this purpose. I just do not want to have always on pc or mac. Smth like pi is ok (nuc?)
And I am fine with cloud services, so I am looking at xeoma now
Hi! I was reading about xeoma and http request now. Could you help me a little more and share a link, where I can read aboud http commands in HA? I googled and found RESTful command but that is all about sending request but not receiving.
Sorry for being noob by the way
Hi!
I had an answer from them yesterday. To send the name using http request you need a script between xeoma and HA . However I have asked for a quote from them for a mqqt client module that can do this in a simple way. When I get it I will order it if they have not doubled the price. This module will then be available to everyone that have a pro licence camera. So please be patient until I know