recognize VS face-api.js

If ARM, see if this is the same issue: https://github.com/nextcloud/recognize/issues/688

Yes, you need to install the Recognize app (https://apps.nextcloud.com/apps/recognize)

When you upload new photos, recognize should automatically schedule jobs to process them. You can check the ocjobs table in your database for recognize related entries and check oc_recognize*queue tables for files that have been queued for processing. Hope that helps. For more questions, feel free to create a discussion on the github repo: https://github.com/nextcloud/recognize

I think this Nextcloud-App is using it: Recognize

For context Im wanting to set it up to use with Recognize in Nextcloud.

I already use recognize in my home NAS. It works fine but it's a NextCloud app (it uses tensorflow through face-api.js). I want something portable and easy to use similar to Google Picasa.

I haven't used them but there are a couple of plugins for NextCloud that do just that. Eg: https://apps.nextcloud.com/apps/recognize

No because doing that on your phone for each and every photo is a bit too much to me, but you should definitely try the server side Recognize app and/or the Face Recognition app! Face Recognition is supported natively in the app and Recognize is supported indirectly via tags.

Sort of...not well. There is an app for NextCloud called Recognize [1].

It has gotten significantly better in the last year or so, but it's still pretty bad and pretty slow.

[1] https://apps.nextcloud.com/apps/recognize

Then I used face-api.js to find the coordinates of each eye.

I already use recognize in my home NAS. It works fine but it's a NextCloud app (it uses tensorflow through face-api.js). I want something portable and easy to use similar to Google Picasa.

https://github.com/justadudewhohacks/face-api.js/ This is what used in web, can't find similar for react native.

For those looking to not only detect, but also vectorize faces (for later comparison against other faces) -- face recognition, try out: face-api.js

https://github.com/justadudewhohacks/face-api.js/

I saw something similar which used P5.js alongside the face-api.js package.

Meanwhile, Lochlann started to work on facial recognition with face-api.js, which was a challenge. Still, the moment he overcame the hurdle and got it working, the team knew it provided many cool opportunities for this project.

For the record - the author considered these two libraries for face detection: https://github.com/vladmandic/human & https://github.com/justadudewhohacks/face-api.js, he chose "vladmandic/human".

I have used https://github.com/justadudewhohacks/face-api.js/ for facial recognition and it works quite well for a self hosted free option.

While I was writing some initial code I found that Vincent Mühler had already created a high-level API called face-api.js built upon TensorFlow.js which wraps all the low-level operations and provides a convenient API for common tasks like face recognition and emotion detection. He also provides lots of example code on his GitHub repository so I was able to build my first emotion recognizer within a matter of minutes.

That's where the problem with your line of thought occurs. I have no relation to that algorithm or its maker(s). It's just a publicly available AI model that a lot of people use.