I’m curious about whether the following is possible. When trying to ID a species, it might be helpful to sort through the (sometimes thousands of) photos to find something similar to mine. Say, a bird is flying directly overhead and I take a picture. Referencing pictures of birds taken from above or the side don’t help as much in this case. If I could easily sort to “ventral view” or something like that, it could be very helpful—no? Having a category of angles (portrait, profile, from above, from below, rear, etc.) could be tagged and then sorted, perhaps.
Would this be generally worth it? If there’s some interest and technical possibility, I can put in a feature request.
This is somewhat related to another topic on sorting photos. Similar sorting of birds by first molt, second molt, breeding plumage, etc., would be helpful. Especially as iNat grows and hundreds of photos for a species turns to tens of thousands.
Same as there, you can create an observation field and start sorting out photos, any field is helpful as long as it is used, doubt anything else would be ever created to do such thing. Adding more views to species photo set can help with that too.
The problem is that such fields will apply to the whole observation, not just the individual photos. There is a feature request to allow annotation/markings on individual photos, such that comments (like “ventral view”, “good diagnostic view”, “male/female in copula” etc) and even drawing directly on photos (such as circling which organism in the photo is the subject) can be done.
But by far the biggest problem, is that someone has to go through and apply all the “lateral”, “ventral” etc… and that effort probably makes just skimming through “all observations of taxa” looking for good comparisons far easier in the long run!
I have a bookmark folder, in which I try to have a bookmarked observation for all the main spiders I have difficulty with, and for those I make the descriptions match the view type that I have saved, to make them easier to find. But I rarely use it… I just browse the taxa observations looking for the right views to compare with, which are often composite (eg oblique anterior-dorsal for comparing an eyefield)
The number of observations (and growth) actually helps, as it raises the sample pool and increases the chances of finding a useful one by quick scanning the observations page. It also means you don’t need to look INTO the observations to see 2nd photo etc, as it is quicker just to bring up an extra page worth of obs to check next batch of 1st photos. Only if you get to the end and still haven’t found the good comparison do you actually need to go back and check 2nd photos etc.
Yes, I figure it would not be very useful unless people added annotations when uploading, or if people could do it on other people’s photos and it would be visible to all.
The trouble with browsing is that there are some cases where it takes a long time. I use the same general method, just looking at many photos until I find what I want. In one case, I wanted a picture of a flying duck and there were hundreds of photos, but none of them had flying ducks. I gave up after having the site load so many pages full of duck pictures. A more detailed sort feature would help.
Also, just looking at all the pictures works okay now, but what about in 5 years when there are an exponential number of photographs? If we started more detailed annotations now, we could better deal with these issues and with new issues that come with more experienced users using the site.
I use local guides myself. They only have so many photos. Sometimes only 1. And photos are easier than descriptions with special birding terminology. Plus, photos are something iNat has in abundance—way better than guides.
My main concern is not myself. I could identify the birds without finding the photo I wanted. I’m just wondering if additional functionality would be practical and helpful to thousands of other people using iNat.
Not everything needs to get to species level ID. Aves is perfectly acceptable!!!
I would just ID to a level that requires an acceptable level of input. Most NZ spiders I will spend a great deal of time trying to identify, only because I want to learn about them. Working through the literature and trying to figure out what the describers back in the 1800s meant is quite interesting! But for plants, most of the time I will just ID the ones I know (not a huge amount), and some I might have a punt if I think there are identifiers that will come after me to correct any mistakes. It really doesn’t matter how fine the IDs get, as long as you enjoy your iNat experience!
And who knows what the future tech will be capable of… we might agonise for hours now over the IDs, only to have the CV advance in the next year to the point where it can come back and ID everything that we haven’t…
Yes, that’s my point: for people who want to learn. That’s a big reason why I am here.
Sure, details aren’t necessary, as others will know. But sometimes they don’t. There’s a percentage of observations that never get to species level, even with clear photos. And anyway, I’m interested in learning and improving the site. Just an idea.
As an analogy, I work for stock photo sites. Some sites are nearing a billion photos. You want a picture of a girl and a dog. Search for girl, dog. You’ll get maybe a thousand photos, but not what you want. You can fine-tune the search to exactly what you want through keywords saved with each photo.
Maybe we can learn and adapt, making useful these techniques and understand how people might use giant photo databases.