iNaturalist Annotation Effort

[Wildlife_Enthusiast]

This post was created so that iNaturalist Annotation Effort members have a place to post their progress on certain taxa, and so that the project members have a place to effectively communicate.

If you are an iNaturalist Annotation Effort member, Only post things that are relevant to the project, We do not need to know about your cat that drowned in a toilet 2 years ago

If you are not an iNaturalist Annotation Effort member, Only post things on this topic if you have a question about the project

Project Guidelines

[ahospers]

To which extent can the ai, computer vision assist in distinguish annotations e.g. “road kill”, “imago”, “Larve”, “Caterpillar”?, “juveniel”, “adult” “male” “female”

(this project will help/ assist in the goal)

[Wildlife_Enthusiast]

Sex annotation : I think CV might be able to do sex annotation in species with high amounts of sexual dimorphism. But with species that have little or no sexual dimorphism i think it would be impossible for the CV to annotate correctly.

Life stage : Most species you can easily do life stage annotations, for example moths is (egg - larva (caterpillar) - moth). With those types of species i think CV could be very useful, as life stage annotations take up time that could be used for other annotations.

Plant phenology : Most plant species that i have encountered are easy to determine if it is flowering, fruiting, flower budding or no evidence of flowering so in these species i think CV could be very useful.

Summary : CV would be useful in “easy” species, but in “harder” species it would quickly become useless.

Note 1 : identifiers would have to “train” the CV with atleast a few hundred observations so that it can reliably annotate.

Note 2 : even if the CV only can annotate easy species, it would save a lot of time for annotators to focus on harder species.

[Wildlife_Enthusiast]

Arge Moth (Apantesis Arge) annotations complete!

The iNaturalist Annotation Effort has finished its first moth species today! the species is called “Arge Moth” and it has 984 observations on iNat.

Results :
Arge moths are most frequently observed in May
There is roughly 10x more adult observations then larva observations
Doris tiger moths are sometimes misidentified as Arge moth

Stats :

Adult : 894
Larva : 85
Misidentifications found : 13
Annotation time : 2 days (This means we finished all Arge moth observations 2 days after we started the species)
User(s) on species : 1

[hkibak]

I’ve been thinking recently about how important and useful annotation is… and I was wondering whether it would be possible to list annotators the way we list observers and identifiers?

[Wildlife_Enthusiast]

From what i know of iNaturalist, there is no way currently to list annotators the way we list observers and identifiers

However, you can put a feature request in if you want.

[dark-shark]

I put it in feature requests.

[dkaposi]

a old feature request on using the CV to annotate:
https://forum.inaturalist.org/t/use-computer-vision-to-annotate-observations/3331

[ahospers]

@Loarie

We could certainly train a computer vision model to predict annotations like life stage rather than species. But remember to train the model, we’ll still need lots of training data so the more life stage annotations you can add the better the model will be

https://forum.inaturalist.org/t/use-computer-vision-to-annotate-observations/3331/20

We could certainly train a computer vision model to predict annotations like life stage rather than species. But remember to train the model, we’ll still need lots of training data so the more life stage annotations you can add the better the model will be

@Tiwane

I hear you. We were talking about your proposal yesterday and it is possible, although it would take some time to train. But as I said in that thread, just as our species ID AI was trained on observations with millions of IDs made manually, so would this AI, and the community would still have to go through and correct AI mistakes for the next training run to improve things. As to whether we have “enough” annotated observations now to train an acceptable model, that’s something we can look into. But more will pretty much always be better.

Bulk Selection of Observations to Add to a Project - Feature Requests - iNaturalist Community Forum

@Alex

First of all, I think we should decouple the feature request for automated annotation from the ongoing vision system. To us, it doesn’t make sense to talk about them together, because we won’t be implementing them together. Happy to talk about that further if you’d like.

With regard to thresholds, the unfortunate answer is that we don’t know what the thresholds are yet. We just know that more data produces better models. It’s taken us a few years to be able to understand and chart the relationship between the number of ID’d photos of species and suggestion accuracy. We haven’t even started looking at how an annotation model would perform, or what reasonable thresholds for one might be. It may be that the thresholds are similar to suggestions, but we don’t know that yet.

https://forum.inaturalist.org/t/use-computer-vision-to-annotate-observations/3331/43

@Loarie

We could certainly train a computer vision model to predict annotations like life stage rather than species. But remember to train the model, we’ll still need lots of training data so the more life stage annotations you can add the better the model will be

https://forum.inaturalist.org/t/use-computer-vision-to-annotate-observations/3331/20

We could certainly train a computer vision model to predict annotations like life stage rather than species. But remember to train the model, we’ll still need lots of training data so the more life stage annotations you can add the better the model will be

@Alex

People submit photos of feathers as evidence and they end up in the training data, but I think a dedicated model just for feathers would probably do better. If you’re interested in trying to create something like that, keep me in the loop! I’d love to know how it goes.

https://forum.inaturalist.org/t/new-computer-vision-model-released/24729/33

[Wildlife_Enthusiast]

Instead of quoting me and editing the words, quote from the right person - The first quote is from a staff member called loarie the second quote is from tiwane.

You can quote from another forum thread and putting it here by copy pasting the quote.

[dianastuder]

see the /20 on hospers link? That goes straight to loarie’s comment.

[cthawley]

There is an existing feature request for this (or at least similar) that can be voted on:
https://forum.inaturalist.org/t/add-annotations-tab-on-profile-page/24964

[Wildlife_Enthusiast]

Changing Project Priorities

• Yes
• No

Should we change the iNaturalist Annotation Effort’s goal from “Annotating entire species” to “Educating people on annotating, misidentifications, etc”

This would mean instead of focusing entirely on annotating, we will try to educate people on how to identify and correct misidentifications in similar species, and we will also try to educate people on the importance of annotating.

We would still be annotating, but the project would be leaning towards education.

We will also rename the project to “iNaturalist Annotation & Education” if project priorities are changed

Edit : Poll is closed, Project priorities have changed and the project is renamed to “iNaturalist Annotation & Education”

[dianastuder]

If you do rename I would spell out
Annotation & Education.
You know what your A and E are, but each new person won’t.
PS you can drop ‘project’ tho.