I agree that they’re worth catching, but I don’t agree that it’s more important than fixing misidentifications. As far as training is concerned, one wrongly identified observation with (say) five photos must surely be worse than a correctly identified observation with one wrong photo mixed in?
2 Likes
Identifiers are valuable people, and I empathize with some of @mikew65’s frustrations as an identifier. But rather than trying to change or restrict the behavior of observers, I’d advocate for identifiers to set their own limits to avoid pain. If you’ve got a slow internet connection, look at 3 or 4 photos, and skip any observations with more. Really, there are just so many more observations that need identifying – you’ll never run out. Skip the hard ones or save them for a better time. Be good to yourself and don’t feel guilty for setting limits. You’re appreciated!
14 Likes
Not necessarily. A misidentified organism often resembles the true organism, and therefore misidentifications often have relatively minor impact on the training process. Otoh, an observation with multiple photos of multiple species can have a major impact. In my experience, it’s not unusual for the lowest common ancestor to be at the class level.
1 Like
Let me try again. Everyone goes through a learning process to become an experienced, productive iNat observer. Training materials have been created for this purpose but unfortunately few users actually use them. Instead the system should offer suggestions in real time. When a new user attempts to submit an observation with ten photos, a prompt should appear. Since the mobile app is currently under development, that’s probably a good place to experiment with in-app training.
2 Likes
But similar species are the ones you’re most wanting to distinguish a given species from. Those sorts of mis-IDs are what get the CV completely up the creek, in a way that turns into a vicious circle. I suspect that one completely different photo amidst a bunch of ‘right’ photos will have no more impact than, say, (a) the distant photo included in an observation with nothing clearly showing the features of the relevant species, or (b) a photo clearly focusing on one species but with a second species in one corner, and using the same photo for two observations - and we include both of them all the time. But if you’re clearly showing similar but slightly different characteristics, that’s downright confusing.
Incidentally, I like the idea of having an ‘exclude this photo from the CV’ option for individual photos to get rid of some of the worst offenders, but it doesn’t look likely at this point.
3 Likes
Slow internet + too many photos = frustration! True! I doubt this will be changed. Skipping over the observations with many photos is probably the best available coping mechanism for now.
(I think the 20 photos limit was ample – did that recently go away? I do hit it occasionally, mostly with bamboo, where I don’t know exactly what angles I’ll need but know I do need multiples.)
2 Likes
I can’t think of any way to justifiably chose which photos are relevant.
If it slows you down, go on to the next one. There are plenty of observations that need ID.
5 Likes
Fat fingers! Happen all the time.
I leave a note and the observer fixes them.
So do I, with a thank you, when someone points out my loading mistake.
1 Like
For an unknown grass I take
- Overall photo
- Flowering head with measurement 1-2 photo, dry
- Flowering head with measurement 1-2 photo, green
- Blade tip, blade back and front
- Node
- Sleeve
- First branch near soil
- Stolon
- Close-up of flowering head, branch of panicle or raceme - different angles
- Close-up of a spikelet
- Close up of glumes
- Close-up of lemmas, back and front, any paleas and grains
This is about 20 details I try to fit into 12 photos.
For an unknown moss,
- Overall dry, from above and from an angle if capsules are present
- Overall wet, from above and from an angle
- Leaves and young leaves, wet and dry.
- Underside when applicable
- Main stem, branch and terminal branch (wet and dry)
- Capsules, multiple photos if different life stages available; exostome/endostome teeth
- Perichaetial and perigonial leaves
- Any deciduous tips
- Rhizoids
- Innovations
- Tomentum, gemmae and tubers
- People with a microscope add 1-10 images
Not all of the above applies to every single moss but I usually end up with 6-8 photos.
20-25 seems to be a reasonable limit.
7 Likes
Unless something goes wrong. A few weeks ago, I created an observation with several photos, didn’t like the blurriness of some, so I took a few more of the same organism. The app would not let me add them; it threw the error message about “only one individual organism per observation” (even though the new photos were following up the same individual) and the one about the limit of 20 photos (both sets combined were less than 20). No matter how many times I acknowledged the message, it kept popping back up. The only workaround was to upload the observation, then go back and edit to add the new photos. I didn’t post a bug report because I figured that – infuriatingly – it was intended behavior.
This sounds like what I just described happening, except that there was no way to get past the prompt. X-ing it out just made it reappear when I again tried adding the new photos. And I’m nowhere near a new user.
Instead of the app I’ve used the web version of iNat on my phone to post an observation or add a photo to an observation from my phone’s photo library. Seems to work.
1 Like
Coming from microscopy, I’ve learned one crucial lesson: always add every shot you have, even if it seems like trash. I often don’t know which key features are needed for certain species or even for their family taxon.
Many times, I’ve had to dig up images I originally discarded due to poor quality or seemingly irrelevant details, only to discover that some of the worst pictures contained the most critical features. I also tend to add multiple images that may look identical, but with different focal points—or in the hope that the animal has turned its head or leg just enough in one of the shots.
You get my point.
That said, I also recognize that non-relevant images can be filtered out later. Sometimes, I review older RG observations and remove images that turned out to be unnecessary for identification after all. I don’t expect that effort from others, though. But, It would be nice^^
6 Likes
Yes, I understand. If you’re reading and/or commenting on this thread, you’re probably not the “new user” I have in mind. A good example is the 1000s of users who will download the app in the next couple of weeks in conjunction with City Nature Challenge. If past years are any indication, the quantity of observations will skyrocket but their quality will be lacking. The only way to reach those people is thru the app they have in their hands at the time they take the photo.
2 Likes
Yeah same here. I’ve only found some here and there.
I sympathise in that our internet is not always stable, but my mind remembers this thread. That was an impassioned plea for more multi-photo Observations.
What I said on that thread begins:
The vast majority of Observations are still one photo, including my own, with the new caveat that sometimes I add an additional photo that I think could be helpful to others, even if I already know what I am observing.
(That said, I am in a newer area so more things are unknown to me now; the new-to-me genus of beetle I saw last night is likely one of three species within the genus that are found in Mexico, all of which look very similar, and I am not at all sure which aspects could differentiate, so I uploaded multiple photos.)
If we want new users to post things to identification level, so that the users are encouraged to develop into “real” users and not just one day participants, multiple angles are likely needed. Photos may look similar, but I remember John Ascher looking at one of my bee Observations and casually noting a leg spur only visible in one of the later photos, a spur I had not even noticed, which is to say I think taxon specialists may be able to see subtle differences that are important because they know where to look.
Maybe I have read too many CNC threads, but most of the complaints I have seen on the forum re: CNCs are (1) that users come for one day only and (2) they leave behind a wave of Observations with only one blurry photo, uncropped, because they are in a hurry to get to the next thing and (3) that these photos are not identifiable beyond a certain point.
I do not recall any complaints regarding a trend for CNC participants to post multi-photo Observations with repeated views, etc, but to be fair, I don’t read all those threads extensively.
The last part of what I said in my comment on that other thread was this:
I still stand behind that, too. And I posted that when I was (fairly) new. I take my cues from the iNatters, old and young alike, whom I respect.
As well as modeling your single-photo preference, you could also put a note on your profile kindly asking people to be mindful of the number of photos they use in Observations in consideration of those with less connectivity. I learn a lot from reading profiles and suspect most newer users do, too.
6 Likes
I’ll sometimes post 3-5 photos of an organism from different angles even when I know the first photo is sufficient for ID and I already know what the species is. This is because I might want to know/remember what that organism looks like from another angle. For example, sparrows. Knowing what the pattern on the back or on top of the head looks like is useful if not crucial in IDing certain species.
5 Likes
Depends how and what you ID.
I see some obs with more than 5 photos, every day.
1 Like
Not-microscopically, I’ve noticed that ‘bad’ photos are occasionally easier to ID than ‘good’ photos. A wonderful front photo of a daisy is gorgeous, but I’ll look at it with a form of grief if it has no accompaniment oftentimes because there’s absolutely nothing I can do with that… meanwhile the blurry crappy distance photo of a daisy next to it I can confidently ID to species at a glance because it happens to include bud shots at the edge of the frame. Blurry bud shots, sure, but for the species I’m thinking of it’d still be unambiguous…
I would like if others would do that review on necessary images sometimes, as long as they’re cautious of it. Some things aren’t necessary to this ID, but may be very useful for my ID of something else, after all - I like to collect references for some very odd details sometimes.
(This of course all varies taxa to taxa. Yet another daisy I’m considering disagreeing to genus on various observations because the only way to distinguish species is some very fussy detailing I know can’t be told at various levels of blurriness…)
3 Likes
Yeah good point. I basically only ID birds.
Yeah I will also usually post multiple angles, at times adding up to more than 5 pictures for a single subject, especially for species that I do not know well. I have a less-than-optimal phone camera that does not always capture all the important features sharply at once, so even though I always make a selection for the best pictures, I don’t want to eliminate details that I might not be aware are important for IDing. I’d rather upload two extra unnecessary pictures and be sure it’s IDable than not upload them and lose those details, even if it might be a barrier to some IDers with slow internet connections.
3 Likes