I don't know houw you would do it on ios but you should be able to do it on android if the phone supports it with.this library from google: https://developers.google.com/ar. Source: about 1 year ago

If you have any control on the choice of the source/webcam, I'd recommend using a camera that can sense depth from the start (lidar cameras, like Intel RealSense if you are building something like a commercial robot; or a consumer device with lidar capabilities like iPad Pros since 2020, because they come with SDKs to do what you want from the start. E.g. https://developer.apple.com/augmented-reality/arkit/ or... Source: over 2 years ago

You guys are right that Unity doesn't support building for arm64 Linux. It looks like the op could potentially install Android on the Raspberry Pi, which may allow them to run Android APKs built with Unity. However, AR Core is needed in order for Unity's AR functionality to work, and I suspect it would take additional work to get AR Core working on the Pi with an external camera and gyroscope. Source: over 2 years ago

If the phone doesn't support ARCore, then you would have to implement all of the world / surface detection yourself inside your application code, which is very difficult problem to solve. Source: over 2 years ago

If you're looking to build a more advanced application, there are plenty of useful resources for all major technologies. For mobile apps, the best places to get started are docs for Google ARCore and Apple ARKit. Both platforms work with popular gaming engines like Unity and Unreal Engine. - Source: dev.to / over 2 years ago

PhotoPrism if you want something local - https://photoprism.app/. Source: over 1 year ago

Not sure I’m a fan of QuMagie - I’ve started using https://photoprism.app with Qfile on the mobile to backup the photos …. Source: over 1 year ago

I recently setup PhotoPrism[0] on my NAS and am happy with it. 0: https://photoprism.app/. - Source: Hacker News / over 1 year ago

I would also recommend checking out PhotoPrism for a robust self hosted endpoint. One of the features that sets that project apart is its implementation of object and facial detection to help classify and sort or search your library. It's worked well for me in limited (<10k imgs) use so far, but I haven't had the time yet to throw my entire 100k+ library at it. https://photoprism.app/. - Source: Hacker News / over 1 year ago

Https://photoprism.app/ if you do like running servers. Source: over 1 year ago