I know there are quite some discussions ongoing on whether or not to include captive species. There seems to be some consensus that it is fine as long as the observation is clearly indicated as non-wild, and that many people actually do use them for research, e.g. in tracking potential invasives, in inventorying species that are used in urban green spaces, or in monitoring wildlife trade. But these are of course still observations of captive specimens where the locality still matters in a way.
There seems to be less clarity on whether these observations are good for the identification algorhythm. I am interested in uploading pictures of captive specimens - partly because it would help me remember and order my material ;-) partly because I would hope it would help the CV recognition. As a biology-professor-cum-amateur-naturalist, I am privileged to visit a lot of zoos, botanical gardens, natural history museums, labs… where I take pictures of pretty rarely seen species that are most likely correctly identified to species level. And I notice that e.g. regarding African plants, fishes and invertebrates, we really need more pictures and identifiers on iNaturalist. Of course, the locality is really not relevant at all in these cases. Would it make sense to consistently upload them?
It’s all up to you and how you spend your time, for species with no photos it makes more sense capturing them! See relevant topics where it was discussed:
https://forum.inaturalist.org/t/who-uses-captive-observations/15354
https://forum.inaturalist.org/t/zoo-animals-why-are-they-permitted/4707
I think we’re not sure how much it’d help with CV - remember, the CV is trained to recognize photos of an organism, not the actual organism itself. So what’s most helpful is to take photos that would be similar to ones people would take in the wild. If the animal is in a zoo and things like bars are in the foreground or the background is not natural, it might not help the CV provide the best suggestions for in situ photos of the same organism in the wild. Similarly, if the CV was trained on a bunch photos of pinned and spread specimens of a certain insect species, it might not be able to properly ID that species from in situ photos where the wings are folded and the background is not white. But, we don’t know that definitively, it would take some experimentation which is unfortunately timely and costly.
Photos of a plant in a botanical garden would probably be a much closer match to in situ wild photos of that plant.
That being said, in general I’d use iNat in ways that make you happy and (in IMO) focus mostly on wild organisms. But the latter isn’t a hard and fast rule, as long as the observations are properly marked.
Thanks for the answers! Good point regarding the difference between recognition of a picture and of an organism itself, I hadn’t thought about that.
On the other hand, I am actually trying to take pictures of as many species as possible (also for use in my classes); I am definitely not a real amateur photographer, but I do try to make decent close-ups in a natural pose etc., also in zoos, and would find it a shame if these pictures stay on my PC only, in case anyone could use them… So then I guess definitely aquarium and botanical garden pictures would probably be useful for CV.
Some friends and I have posted cultivated individuals of species that are beginning to escape and invade wild lands, and species that may do so soon. I think that’s potentially useful. Also, I posted a bunch of cultivated plants with the flowers dissected for use in a class. Training the CV could be another use for these things, though @tiwane makes a good point about bars – or pots!
I’d like to see “Cultivate/Captive” removed from “Casual” (a.k.a. rejects). Right now, these potentially useful records are mixed in with observations lacking dates or places, or unidentifiable. I’d like to see three categories (normal needs ID and Research Grade; Captive/cultivated I and II [without or with community ID]; and Casual [something’s wrong]).
Aha ok, you are pointing me to another potential use of iNaturalist images: didactics, even of dissected specimens… So in that sense we could also see iNaturalist as a forum to share high-quality (well-identified, diagnostic structures visible) pictures of organisms? Because this is mainly why I take so many pictures of living organisms wherever I go (a.o. for my own teaching), I’d find it great to share…
I think that is a good use of iNaturalist!
In this same vein, if you make the license to your photos a fairly permissible one (like CC0 or CC BY SA) those photos of rare species can be used by Wikimedia which makes them more widely usable by lots of people.
I identify a lot of cultivated plants. I know there is a concern that photos of captive/cultivated species might be different from photos of them in their natural habitat. In my personal opinion, this isn’t that big of a concern - it seems to me that the CV still does a good job recognizing plants in any setting (e.g. Fittonia albivenis). In fact, for many outdoor plants it can be kind of hard to tell whether the plant is cultivated or wild based on the photo alone, and the location helps a lot - whether the pin is in the middle of a suburban neighborhood, or a national park, etc. I think your photos would be helpful!
I think such photos could be quite helpful, especially for species that have few or no observations yet. Even better if you can capture photos of less often observed plant states, like seeds (it’s amazing how few species actually have photographs of their seeds on inat).
@tiwane This makes me wonder though, how much do captive observations influence the CV? Such enormous numbers of the captive/casual observations are misidentified. Should identifiers be putting time into correcting that stuff for the sake of CV?
I like to document captive plants that are growing in places far from their native range. In some cases it has proven useful in tracing where an invasive plant has started spreading from, and in others it can be useful for horticulturalists who want to know in which climates a specific plant can be successfully grown. Either way, I do think there is value to it. Especially for organisms that are infrequently seen!
At heart, iNat’s about sharing and discussing one’s observations of wild organisms. I think it’d be much better to post the photos to Wikimedia and they can then be used as taxon photos on iNat.
I’m not a CV dev (or a dev at all!) but so far, we don’t filter photos by captive or non-captive when training the CV model - training photos have basically been chosen at random. So I don’t think we can really know how many captive vs. non-captive photos were used for each taxon. We’re doing experiments with how to better train the model, though, so all of this might change.
The sake of the CV is not the only reason.
If only one person has tentatively IDed that plant as maybe This plant then iNat will use that to suggest … Seen Nearby - which makes it worthwhile to clear the errors we come across
Hmm. If CV uses cultivated/captive observations for training, that’s yet another reason to separate the captive/wild dichotomy from the Needs ID vs. not dichotomy! Come on, iNaturalist!
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.