[FEAT] Local labeling by non-dev end users

[ivelin]

trafficstars

First of all, this is an awesome project! Congrats!

Is your feature request related to a problem? Please describe. In its current form labeling requires some development skills to setup the web app. Question: Is it a goal for the project to eventually allow non-tech end-users to apply in their homes?

Describe the solution you'd like If the answer to the question above is no, please close this issue. Otherwise, it would be nice to think about a way of delivering the core of this project to non-tech DIY users so they can use and train a private face id model in their household.

Additional context We've done some work in this direction with ambianic.ai. It targets user who can burn an SD card, assemble and boot a Raspberry Pi, but are not comfortable with ssh and code. Happy to chat more about integration if there is interest.

[ozett]

interesting link to ambianic.ai. https://blog.ambianic.ai/2021/09/02/movenet-vs-posenet-person-fall-detection.html

may you introduce your efforts to the frigate-ideas of implementing a custom model?

1: building a model https://github.com/blakeblackshear/frigate/discussions/1426

2 : general discussion https://github.com/blakeblackshear/frigate/discussions/1043

[ivelin]

Thank you for the links @ozett . Done: https://github.com/blakeblackshear/frigate/discussions/1043#discussioncomment-1525002