Platform(s), such as mobile, website, API, other: Backend URLs (aka web addresses) of any pages, if relevant: - Description of need: This change is needed to reduce the huge workload for identifiers who have to clean up wrong computer vision based IDs Feature request details: A few years back, …

[image] steven_molteno: We’re not keeping up with the corrections (it’s grown to 2000 misidentifications, and more every month) It would be good to see hybrids like this included in the CV suggestions. Another example would be Crocosmia, where most plants observed are hybrids. But remember y…

Possibly, many more photos are needed for reliable hybrid computer vision models. For Gazania × splendens, there are currently 3,590 research grade observations, representing roughly 5,000 photos. It could be interesting to train a model that includes hybrids for which at least 5,000 photos are ava…

@tiwane We really need an update on this

Hibiscus rosa-sinensis is now Hibiscus x rosa-sinensis in iNat. I suspect that some of the plants grown as Hibiscus rosa-sinensis are further hybrids involving other members of Lilibiscus. (I’m also surprised to find as many as 11,500 observations of Hibiscus x archeri compared to 70,000 observation…

To add to the list from a British viewpoint Symphytum x uplandicum is the commonest comfrey in Britain. Symphytum x hidcotense is comparably common to Symphytum grandiflorum. Silene x hampeana occurs frequently. Stachys x ambigua Leucanthemum x superbum Malva x clementii is commoner than it’s pare…

[image] rudolphous: Possibly, many more photos are needed for reliable hybrid computer vision models. For Gazania × splendens, there are currently 3,590 research grade observations, representing roughly 5,000 photos. I understand the motivation to speculate on possible barriers to including h…

The only substantial barrier to this that I can see is indeed a total lack of interest from staff. If that’s not the case, and any of the staff have an interest in reenabling the Computer Vision for nothotaxa, I hope we will see the fruits of that soon. Otherwise, it seems that turning off the CV fo…

[image] rupertclayton: But there was a problem with Mallards and Wood Ducks and so all hybrids were thrown out with the waterfowl bathwater. I’m not even sure that the problem with hybrids affected birds in general rather than just a couple of common and sexually promiscuous waterfowl species…

I always had the impression that plant hybrids operated well in CV, but it was only the bird ones that were controversial (like mallard hybrids). Some of the most common plants out there like Montbretia and Gazania are nothotaxa, meaning they are excluded from CV, even though they have a considerab…

Make computer vision include hybrid taxa on an opt-in basis

Feature Requests

cyanfox August 26, 2026, 5:30am 74

https://forum.inaturalist.org/t/recommendations-on-improving-the-ai-algorithm/63027/35?u=cyanfox

Topic		Replies	Views
When could hybrids could be included in the CV? General	15	1275	September 4, 2023
New Computer Vision model released! News and Updates	42	6324	October 4, 2024
Recommendations on improving the AI algorithm? General	47	1154	April 13, 2025
CV suggesting some hybrid taxa, but hybrid taxon pages do not display "Computer vision model: included" badge Bug Reports	7	290	July 8, 2024
Suggestion: Allow hybrids that specify a genus only Curators	25	2868	August 26, 2026

Make computer vision include hybrid taxa on an opt-in basis

Related topics