Currently, the CV only suggests species for the first image in an observation. It would be useful to be able to get suggestions on any image in an observation. There are ways to achieve this in the current system, but they are cumbersome and only work on a user’s own observations, not those of other…

It already selects pictures completely at random for the training, there is no question of using only the first picture; if there was any potentially viable change to the selection system it would probably have to be to deliberately choose training pictures in a more diverse way than random, not les…

It should not be hard at all. It actually already works like that in the mobile app. Only the web version looks just at the first photo.

If I’m understanding this as a reponse to what I said just above your comment, it works like this in Android (no luck on iPhone?), which is part of what makes mobile IDing ironically much easier than desktop in some ways, alongside SIGNIFICANTLY faster loading times between pictures and observations…

[image] wildskyflower: It already selects pictures completely at random for the training, there is no question of using only the first picture Is this actually true? I had the impression training used all photos. [image] wildskyflower: Of course, training the model to actually score all …

[image] yerbasanta: Is this actually true? I had the impression training used all photos. We train on all photos for taxon where there are one thousand photos or less for that taxon. If there are more than one thousand photos for the taxon, then not all are used.

[image] tiwane: We train on all photos for taxon where there are one thousand photos or less for that taxon. If there are more than one thousand photos for the taxon, then not all are used. I wonder about this approach a little, at least when it comes to trees, people tend to upload the 1st p…

Most machine learning models use only a subset of available data as training data. Using all photos to train would not be computationally feasible (some taxa have millions of photos). For training models, there’s a point at which adding additional training data is not worth it - the cost of the addi…

Using CV on each photo could also help point out observations of different subjects. Right now there can be a taxon in the first photo but a completely different taxon in other photos. CV could send a warning if there is a taxon mismatch detected among different photos of the same observation.

Just to note: you can now do this in iNaturalist Next

it has always been possible on the android app

Use computer vision on each photo in an observation

Feature Requests

bouteloua February 16, 2025, 12:23am 46

see https://forum.inaturalist.org/t/inaturalist-next-app-soft-launch-for-iphones/55277

Topic		Replies	Views
Workings of iNat Artificial Intelligence (AI) aka Computer Vision (CV) General	28	2661	March 16, 2022
Some 'interesting' AI suggestions of late General	41	2254	September 10, 2022
CV suggestions are bizarre, not "visually similar" at all Bug Reports	65	2236	February 2, 2023
"Helping" the computer vision - is this wrong? General	36	2881	July 12, 2020
Did computer vision get worse lately? General	33	3278	April 13, 2021

Use computer vision on each photo in an observation

Related topics