I am facing a lot of trouble recognizing the Bird Chicks in absence of their Parents. In most cases they don’t look like their parents at all at their early childhood. Can the problem be resolved by the system? Any suggestions?
P.S: I tried to get an id uploading few such images but failed to get any reasonable system suggestion.
Not surprising. Bird chicks don’t resemble adults much at all, so the iNat CV likely won’t recognize them unless there’s a bunch of pics of them already in the system.
3 Likes
I’m not surprised because I know that. I would be very happy if this issue can be solved by iNAT.
The solution is more photos on iNat of chicks that are confirmed to species. The computer vision needs that info to propose IDs.
3 Likes
The problem is how can the confirmed photos be uploaded? Who will do that? I tried this by uploading one such Photo of Asian Openbill Chicks along with other Photos of them with the Adults with Id Suggestion. That was not confirmed by anybody. So for good reason that picture will not be on the CV/system as any reference.
Where possible, make the first picture of your observation one that shows the adult with the chicks - that will help others to confirm more easily. Also remember that all the photos for an observation must be from the same date. If you have photos from different dates, post them as separate observations. You can link multiple photos of the same brood by putting links in the Notes or with fields such as “Similar Observation Set”.
3 Likes
That will not help to identify the chicks. Confirming the Adult is easy and my objective was not that. I know if I upload a typical photo of an Adult Asian Openbill with Cormorant chicks with him/her (it can be seen very commonly in our country), people will mark the photo as that of an Openbill.and all other photos and aspects will be ignored.
The other part of ur comment is a separate issue. I never post photos of species of different dates and locations within a single observation. I upload two-three photos taken from different angle duly mentioning date and location so as to help identification by others. Thanks for ur reply.
Yes, it will. Once your observation has been confirmed by the community, all photos in it can be used to train the computer vision algorithm, including photos of chicks only. If enough people add observations with photos of chicks of different species, the computer vision will start to offer the correct suggestions.
If you were to upload observations with photos of chicks only, it is less likely that the community will confirm your ID. Even putting photos of the chicks first, as you did in the observation you referred to, makes it less likely you will get a human ID. I was able to confirm your ID by looking at the other photos, but be aware that identifiers by default are presented with the first photo, and may skip over observations they are not confident in identifying.
You might also consider writing a journal entry, illustrated with relevant photos, to help teach others how to identify chicks, if you already know how to do this. Then you can link to it on observations and help create a community of identifiers who can do these IDs.
5 Likes
Thanks for explaining the issue. Yes, now I got your logic and will upload images of chicks of birds accordingly.
As I’m a very new entrant in this forum, Journal writing etc was not known to me. Thanks for informing me of the same. I’ll surely try it.
3 Likes
Conclusion: Bird Chicks will remain as Unidentified Fellow to iNAT for ever! They’ve to wait to get a name!.
Adding more identifiable observations with photos of chicks is the key step. You don’t need to write a journal post. @deboas just suggested it as a way to assist others in adding correct IDs to more observations of chicks.
2 Likes
No, No, I got every valuable points suggested by @deboas.
Journal writing will take time and that will not train Computer vision. That’s a separate issue.
Regarding identification of chicks by the system, I duly noted the points as suggested by @deboas that more and more photos of chicks are to be uploaded by observers in the way as advised by @deboas. In fact, I’ve then uploaded one such observation of the Grey/Spot-billed Pelican with chicks although not yet confirmed by anybody yet and hence is valueless till now.
While going through this outstanding Site, I felt that inability to identify/suggest any bird species Id for the bird chicks is a shortcoming of iNAT which may have to be/can be resolved.
When I uploaded the Image of the Openbill Chick, I got Top CV suggestion as "Wild Boar, Barn Owl, House Shrew’, none of which are a even a nearly similar species.
So I posted the issue in the forum and got the valuable suggestion.
The problem is most Birds don’t breed everywhere and uploading the images of chicks may not be possible by everybody. For Example, Spot-billed Pelicans breed only in some parts of South Asia etc. I’ll not be able to upload the images of maximum Wild Duck Chicks as they don’t breed in our Country. Hence, I’ll like to bring the issue to a broader audience and request everybody to upload images of bird chicks following the suggestion of @deboas, so that they can be identified by iNAT CV. A very few person’s effort will not help.
The topic failed to attract sufficient attention of viewer, chances of uploading of Bird Chick images are very lean. Hence I commented that in my previous post.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.