Face clustering: 42 faces left

[31337-4554551n]

Seeing an issue where 42 faces are never being processed.

Unsure if separate or related issue, as new photos are taken, most are being recognised, but in a few cases (maybe 1/5) the face isn't being detected. This is selfies/photos with obvious people.

Original discussion:

Face recognition: 0 Queued files, Last classification: 12 hours ago, Scheduled background jobs: 0, 

 Face clustering: 42 faces left to cluster, Last clustering run: 12 hours ago, Scheduled background jobs: 0,
A minimum of 120 faces per user is necessary for clustering to kick in

If I'm reading this correctly, I have 78 pics with faces detected, but they won't be added into their correct people until I take another 42 photos with faces in them?

Is there any way to shrink those batches? I definitely know I have photos with peoples faces that aren't showing up under people... and idk how to force em

[edit] to clarify I am using recognize to feed into memories

Originally posted by @31337-4554551n in https://github.com/nextcloud/recognize/discussions/1167

[github-actions[bot]]

Hello :wave:

Thank you for taking the time to open this issue with recognize. I know it's frustrating when software causes problems. You have made the right choice to come here and open an issue to make sure your problem gets looked at and if possible solved. I try to answer all issues and if possible fix all bugs here, but it sometimes takes a while until I get to it. Until then, please be patient. Note also that GitHub is a place where people meet to make software better together. Nobody here is under any obligation to help you, solve your problems or deliver on any expectations or demands you may have, but if enough people come together we can collaborate to make this software better. For everyone. Thus, if you can, you could also look at other issues to see whether you can help other people with your knowledge and experience. If you have coding experience it would also be awesome if you could step up to dive into the code and try to fix the odd bug yourself. Everyone will be thankful for extra helping hands! One last word: If you feel, at any point, like you need to vent, this is not the place for it; you can go to the forum, to twitter or somewhere else. But this is a technical issue tracker, so please make sure to focus on the tech and keep your opinions to yourself. (Also see our Code of Conduct. Really.)

I look forward to working with you on this issue Cheers :blue_heart:

[marcelklehr]

but in a few cases (maybe 1/5) the face isn't being detected

This could be due to using WASM mode, or because of shortcomings of the face detection model. It's not easily fixable.

Unless the missing faces are in the 42 faces that are left.

How many pictures do you have overall?

[31337-4554551n]

I am using a WASM module, this is on the linuxserver.io docker, it's alpine based, and I spent a long time trying to get it to work properly, but it's just missing too many dependencies for GNU lib C, there's a github with it compiled, but then it complains about missing nodeJS and other stuff.

A little under 10,000 photos on one account, and a little under 9000 on another.

So the issue with new photos being taknen and a few here and there not showing up probably applies to the stuff I backlogged when I first set it up, but it's much harder to cross reference

[marcelklehr]

A little under 10,000 photos on one account, and a little under 9000 on another.

This is peculiar. With so many photos there should be enough detected faces to sample in enough to get over the 120 faces limits. Except... how many people do you have? If you have all detections grouped into very few people this might explain why it's not overcoming the sample limit.

[31337-4554551n]

It's not detecting the 42. And it's not detecting existing faces, like I'll go out, take 3 photos of the gf,and it'll import 2 for example. There's be maybe 20-30 faces. There were more, but I deleted them when it picks out like random people in a crowd or something.

[marcelklehr]

20-30 faces or 20-30 face clusters (people)?

[31337-4554551n]

20-30 clusters. 1000-2000 pictures of faces.

[31337-4554551n]

What more information is being waited on? I gave the last response

20-30 clusters. 1000-2000 pictures of faces.

[marcelklehr]

Yes, I'm sorry, I didn't remove the "waiting for more information" tag