I’d like to suggest a limit on the number of relevant photos for an observation. Sometimes in a submission there are too many to go through for a single species (record) and many do not add to the overall ability to assist identification anyway -they can be duplicates. If one has a tenuous internet link (regional) it can be rather time-consuming to go through identifications for other people. It is especially frustrating if, like myself, one likes to go through all images to ensure they are all of the single/same species. Personally, I think this might encourage more users to identify.
Your idea will probably not be agreed with by most users of the platform, though I do understand your annoyance, I too have slow internet.
A feature request for a new url filter that sorts by number of pictures attached for the explore/identify pages would probably gain more traction.
4 Likes
Personally, I think you’re trying to limit other observers to conform to your personal limitations.
You can choose how many of the photos you want or need to see to make an id or not.
5 Likes
The current limit is 20 photos per obser.
13 Likes
My thinking on this, which might be wrong, is that naive observers might capture a key feature by accident, and that this might help the computer model train. So this makes me think that we should encourage multiple photos, so long as users upload them correctly (as one observation/place/time) and each photo is slightly different.
10 Likes
That does not sound like a bad idea, at all -it would allow one (especially those with slow internet) to move more smoothly through the identification process. As it is now, increasing the size of an image to aid with identification can take an unnecessary long time and if there are lots of images then… My main point is that some posts have -in my opinion- too many duplicate images that do not help identifiers but slow the process down.
1 Like
Of-course I can choose how many of the photos I need to see to confirm an identity for myself, however I like to do the right thing by the process and check everything. But I have come across numerous instances where an accidental image of another altogether different species has made it into the submission and I was wondering whether this might be something we can do something about.
1 Like
That’s not what I read. You hadn’t brought up the issue of more than one species
That can be dealt with in the DQA
1 Like
That’s part of the problem, in my experience, and as I see it, that many of the photos can be the same. (Not only that, as I mentioned above, there are sometimes images among them of completely different organisms.) I guess, without the access to filters limiting the number for myself, it may be something I’ll just have to tolerate.
Is this an issue with app users mostly? I could see someone taking a dozen or more shots of something on their phone and, instead of reviewing the pics to see which one or two are in focus and best capture the subject, they just submit all of them. But I’ve seen others submit many pics from multiple angles which isn’t a bad idea to capture helpful characteristics.
3 Likes
Maybe that’s the explanation, not reviewing the images beforehand when using their phones.
If the many images with multiple angles help then that is a good thing, of-course.
1 Like
You should be able to create an observations with more than 20 images. This restriction is a few years old, there are older observations with way more than that. It’s also possible to add more images to an observation that has 20 images. Personally I rarely come across observations with more than 3-5 images.
11 Likes
There are two main reasons why I will fight for the ability of users to include at least medium amount of photos per observation. First reason being that newer users or even experienced users encountering a particularly rambunctious field specimen (insects, looking at you) may capture key features in focus in a variety of shots. It may be 8 blurry shots of the same beetle, but a clear tarsi, clear dorsal, clear underside, or similar may be the difference between a family ID and a species ID. The second reason being that as somebody with moderate knowledge of which features may be needed for ID on specimens, the number of features needed for ID on a specimen may be upwards of 7 different angles. Elytra, clypeus, prosternum, pronotum (dorsal and ventral), maxillary palps, abdominal tergites, and tarsi may all be needed in clear microscopic focus to distinguish between species. These cannot be obtained in fewer than 5 shots, and while it may not need 20, things like a whole-body dorsal, an in-situ habitat photo, and size reference are also helpful and I’d rather have all of this information up front than try to chase down somebody hoping that they collected the specimen and can get these additional angles.
So generally, it is a net positive to have a decent number of photos of the same specimen, even if sometimes it results in lumped observations with multiple individuals some of the time.
18 Likes
Yes, many insects require multiple angles for ID. I spend more time wishing that users had gotten additional or different angles than I do wishing they had included fewer photos.
Another factor here is that if one is not an expert in the taxon one has photographed, one may not know what key features are needed for ID. So those of us who frequently observe difficult-to-ID insect groups may post all the images we have that are reasonably in focus because we don’t know which ones will be useful. After a while one does generally develop a sense of which taxa are tricky and which ones can be ID’d with just a clear dorsal photo, but every so often there will turn out to be surprises, like the unexpected out-of-range specimen where one regrets not having taken more photos, so on the whole, I think there are advantages to erring on the side of more photos than are strictly necessary.
Another reason I may include more photos than are needed for ID is if I want to document an interesting interaction or behavior.
As an IDer, I don’t generally mind looking at extra photos (though in most cases I do think the benefits of additional photos tend to diminish after about 5). Occasionally there will turn out to be interesting finds, such as a second species when the observer thought they had only seen one – if it is an old observation and the observer is no longer active/responsive, this is frustrating, but in many cases they are quite happy to edit.
16 Likes
Because of that, more is better.
We have the DQA for - these pictures are not all of a single subject - but the human still has to look thru them.
Identifiers would appreciate if observers made the effort to sort thru their photos and ditch the junk (crop as needed), leave a note ‘for the beetle’. Hopefully will be part of the better onboarding.
I do a lot of identifying, and have learnt to decide whether to give this obs my best shot. Or Mark as Reviewed and move on. If I see ‘I will just empty my memory card from today into this obs’ I skim to confirm at least 2 different subjects - leave a text expander comment and apply the DQA. Which makes it Casual and takes it out of Needs ID for future identifiers.
8 Likes
Yes, it can be dealt with - the point is that the OP feels the need to review every photo of every observation to ensure that there is only one species, to deal with the problem. I’ve definitely been guilty of not checking and having others point out later that there’s one rogue photo, and I’ve also found rogue photos in observations that previous identifiers haven’t noticed. It’s true that the more photos, the less likely that such photos will be noticed, and I sympathise with the slow internet issue - but I doubt it will be changing.
Incidentally, in the initial post:
Clearly, more photos means, more angles… means more accurate identification. That makes sense. The photos that I refer to are those multiples that all look exactly the same, with little or no difference, like the ones I often come across with birds, as that is the group I deal with the vast majority of the time. I understand the need for insects and others.
Exactly. From a training perspective, catching one of these observations is much more valuable that finding a simple misidentification. I wish there were tools to help find these observations.
Some observations have too many photos, that is very true. Some observations have too few photos, that is also true. The system should detect such situations and offer the new user some suggestions. The new mobile phone app would be a good place to implement such a real-time training system.
You are maybe suggesting new users should be iPhone users that prefer to upload from the app?