Add observation - Automatically tagged my observations with incorrect names based on photo filenames

Yep. I’m uploading images right now with the file name Lion Rock, or Lone Kauri Road, (these are the location names where the observations were made), and as the image metadata loads, iNat automatically defaults the species ID to the big cat ‘Lion’ (even though these are rock pool obs), or the tree ‘Kauri’, (bird obs). Seems to be choosing the ID from the photo file name rather than the image.

2 Likes

https://www.inaturalist.org/taxa/51392-Morone-saxatilis had a common name of “rock” in English which was the source of those misidentifications. That name seems absurd so I deleted it (having grown up in an anglophonic Atlantic place and now living in an anglophonic Pacific place, I have never heard anyone refer to this species as a “rock”). Issues like “Lion” and “Kauri” are harder to resolve since they are way more legit common names. FWIW, a cursory glance at the file names associated with iNat photos suggests:

  1. Most people do not rename their photo files
  2. Among people who do, using species names is WAY more common than place names
  3. There are, of course, some delightful people who do both

So I’m inclined to believe the current treatment of helping the species-namers while mildly inconveniencing the place-namers is doing more good than harm. Personally, I think you’re all crazy. Doesn’t it take enough time out of the day just to add these things as keywords?!

3 Likes

Takes no time at all: My phone’s camera has an option to add location name to the file name when taking a photo. the names are woefully bad, so I turned the feautre off. It can also add weather conditions to the file name, imagine that! Maybe it will be useful one day if iNat adds meterological observations.

3 Likes

My images are used for other reasons, not just iNaturalist, and it’s helpful to the receivers of any images I send, to see location information in the file name, particularly when I send a variety. Also with my filing system, which is by species name, it helps to quickly alpha search by location (from the file name in that species folder), if I need to send an image of a species from a particular location.

I’m not bothered by the random auto IDs, more amused than anything when iNat suggest Lion for a sea anemone. At least now I know why it’s happening, mystery solved. BTW, some would say I’m wacky, but maybe not crazy ;-)

7 Likes

Maybe it should only auto-detect it if a) it detects it all the way down to species or b) the user has opted-in somehow?

4 Likes

@danaleeling, for a person to label their photo with the correct species name, the person has to know what they’re looking at. Many of us do not know the genus or even family, etc. of the organisms we are documenting.

4 Likes

I keep having problems with the automatic labelling of species with words pulled from the names of photo files as well, mostly when I label something as “orange bug” or “orange mushroom,” in which case it gets labelled as Citrus x aurantium (=orange, the fruit). But the oddest one I’ve seen so far was this:

Somehow there’s a hash table that defines “skunk” not as Mephitis mephitis or its relatives but as a synonym for cannabis. (Does “weed” in file names also yield species labelled as Cannabis sativa?) Anyway, if the AI drew from both the name of the file and the content of the image, this problem would probably not exist – at least not in the same way.

12 Likes

That’s really funny!

1 Like

From: https://www.inaturalist.org/taxa/75997-Cannabis-sativa


“skunk” is a listed common name.

There are a number threads on this topic of common names including:
https://forum.inaturalist.org/t/reliable-sources-for-common-names-on-inaturalist/5579
https://forum.inaturalist.org/t/erroneous-and-uncommon-common-names-as-the-default/7151
https://forum.inaturalist.org/t/which-common-name-to-use-the-most-common-or-taxonomically-accurate/5952
https://forum.inaturalist.org/t/fixing-incorrect-uncommon-common-names/5739
and many more. My first reaction would be that a slang term such as “skunk” which denotes another species should not be listed as a common name. The common name really should be, for a plant, a botanically used common name, I would have a hard time believing that skunk would be considered a botanically appropriate common name.

4 Likes

Yeah when I search just “skunk” it comes up as the first option…

3 Likes

There seems to be a recent influx of seemingly random observations labeled as the genus Aa (yes, that is a real taxon name). At first I thought it was an issue with the computer vision or misidentification on the observers part, but it seems unlikely that that is the case given the content of the observations (ranging from lichen to something that I think is a moth) and the fact that they seem to be coming from fairly experienced users. I suspect that the issue is due to some kind of alphabetical sorting system that is part of the taxa list or something (I don’t really know how the system works so I am just guessing here).

5 Likes

Any URLs of these observations that you can share?

2 Likes

https://www.inaturalist.org/observations?taxon_id=576873

https://www.inaturalist.org/observations/36519916
https://www.inaturalist.org/observations/36517477

I was just going to post about this issue as well. I’ve seen several of these faulty IDs coming from the same users

1 Like

As a side note, I also noted other observations by these users identified as “Bb” or as something completely wrong (off the top of my head, there was a tree observation that was ID’d as salamander chytrid)
While I’m sure the former is definitely related to the “Aa” issue, I have no clue if the latter is as well

1 Like

Another way to see some of them:
https://www.inaturalist.org/observations?place_id=any&subview=grid&taxon_id=576873

Could this be a result of the photograph’s original file name having an “aa” in the file name? Or a “bb”?

3 Likes

Aside from the incorrect "Aa"s from a user the other day, I also saw an observation or two ID’d as “Bb” (for which I offered more sensible IDs). However, trying to search now on “Bb”, that two-letter combination does not appear to be available in the taxon pick list, so I wonder if it was a temporary glitch that allowed it to even be accepted then? Only the abbreviations for bird common names come up now when “Bb” is typed in, although none of these has only two letters so there is no actual match.

1 Like

It looks like the filename parser is matching against all taxon names. Anil’s Bush Frog has a deprecated name “Nov”, so the filename parser is probably guessing based on the month abbreviation. Next month it might decide everything is a Decapod, unless it only accepts exact matches.

Other three-letter month abbreviations which may cause trouble in the future:

  • Mar - the Zazaki word for snake
  • May - several things, including a genus
  • Jun - the Tseltal word for Ficus insipida

Edit: see https://forum.inaturalist.org/t/extract-species-name-from-photo-filename-upon-upload/6510/14
Exact matches on scientific name only. So only “Nov” and “May” are going to cause misidentifications.

4 Likes

Wondering why when I upload batches of images they all “auto-ID” as “Domestic Horse” before I even get a chance to enter ID’s.

It might have something to do with naming files. Today all my files were named like this: 'Eastern Skunk Cabbage in Intermittent Creek off of Horse Trail #19 a 12-27-19.jpg"

Perhaps iNat is defaulting to the word “horse” in the file name???

Ken

image|690x389

2 Likes

Yes it is a new feature where the site attempts to classify a species based on the contents of the file name. It’s been buggy to say the least, and is already documented.

2 Likes

Thank you! I hope it gets fixed fast :)

2 Likes