Nutcracker
Stop Brexit!
Yep, Brambling is correct! March this year, in Northumbs.
But I hope Pine Warbler wasn't its second choice??!
-
Welcome to BirdForum, the internet's largest birding community with thousands of members from all over the world. The forums are dedicated to wild birds, birding, binoculars and equipment and all that goes with it.
Please register for an account to take part in the discussions in the forum, post your pictures in the gallery and more.
Auto-ID of birds in photos—now for UK and Ireland (2 Viewers)
- Thread starter idaviesbird
- Start date
Yep, Brambling is correct! March this year, in Northumbs.
But I hope Pine Warbler wasn't its second choice??!
Paul Chapman
Well-known member
Very interesting. Why isn't Pine Warbler the second choice on the basis of the shape of the wingbars? The computer will analyse on features.
I 'assumed' like Fugl that it was a Chaffinch and only when Mark highlighted that the computer called it a Brambling did I revisit the photo. Anyone else prepared to admit that they did the same?
All the best
stuartvine
Well-known member
:t: Yes
Now I'm impressed.....
Everything apart from xbills would be a good test, I think we are struggling with crossbills to reliably identify them to start with from photos so 'garbage in, garbage out' unless the images are vouchered with DNA and sonograms....
Anyone want to put it to the test? They don't have a huge number of images available for Europe, unlike North America where the program really excels. If more folks put images into eBird checklists over here...
Nutcracker
Stop Brexit!
I'll have to look to see if I can find any other crap pics that I've not got round to deleting yet 3
Scottish Crossbill is impossible even with DNA (not distinguishable from other European crossbills) and sonograms (Harvie-Brown neglected to get a sonogram of the one he shot in 1870 which became the type specimen of Loxia scotica, so its calls are unknown :eek!
I tried with warblers, good identifiable photos. It got right four out of five Chiffchaffs, but its first choice of the fifth was Willow Warbler. Of Willow Warblers, it similarly got right four out of five, the fifth it thought to be Marsh Warbler, and the correct answer was only the third choice.
It identified Icterine Warbler correctly, but had problem with Eurasian Reed - first of them it identified as Savi's Warbler and second Common Nightingale.
Quite impressive, I think.
It identified Icterine Warbler correctly, but had problem with Eurasian Reed - first of them it identified as Savi's Warbler and second Common Nightingale.
Quite impressive, I think.
Yes it seems so.
The first choice of the attached photo, taken at Linn of Dee, Braemar, Scotland, was Common Crossbill and second was Scottish Crossbill.
Attachments
![Loxia_20090225_Linn_of_Dee_LH6D8670[1].jpg](/data/attachments/592/592533-9935069fabd8c6c57445e83dcb56807a.jpg)
Nutcracker
Stop Brexit!
Worrying though if it did that in Britain as well as in Finland, identifying a common bird as a decidedly rare one. One feels that with artificial intelligence like this, it should be set with known mapped distributions to favour common over rare where it has any degree of uncertainty. Better to miss recording one or two genuine rarities, than get lots of false positives for rarities that aren't there.
Yes it seems so.
The first choice of the attached photo, taken at Linn of Dee, Braemar, Scotland, was Common Crossbill and second was Scottish Crossbill.
I'd be strongly tempted to call that one a Parrot Crossbill! What a massive neck and bill :t:
Mike Crawley
Emeritus President at Burnage Rugby Club
Downloaded the app. yesterday and it has correctly identified the photographs of birds I have taken - all common stuff but quite a few are a little obscured by branches/feeders, etc.
However, I downloaded a picture (fabulously clear) of a Wilson's Storm Petrel (completel with yellow webbing) and it didn't have a clue. Maybe an unfair test ?
However, I downloaded a picture (fabulously clear) of a Wilson's Storm Petrel (completel with yellow webbing) and it didn't have a clue. Maybe an unfair test ?
Last edited:
So it does tempter its decision-making if you tell it where you are.... you can either ID with zero date location info or enter date/location this means that if the difference isn't strong it will opt for the locally common species based on available eBird data....
However, when I played with this it was less likely to call difficult taxa like Wilson's Snipe vs Common Snipe correctly. It was erring on the side of caution despite apparently getting these species when not being told where it was...
AFAIK it won't know Parrotbill, Thrush Nightingale etc yet anyway as the Europe pack hasn't been released. Even when it does it still likely needs more pics to get really good... Likewise with Acros etc....
which pic?
McMadd
You should see the other bloke...
Are they not a little wary of the T&C if you do so...?
Mike Crawley
Emeritus President at Burnage Rugby Club
I borrowed one from Joe Pender obtained off a Scilly Pelagic
http://wwwsapphirepelagics.blogspot.co.uk
McMadd
You should see the other bloke...
Well I'm never in favour of "free" unless it's in my favour...I recall some comment somewhere in BF or elsewhere (not searching for it) disagreeing quite strongly with those T&C for reasons unknown...never bothered looking for them so you saved me a job there...
Paul Chapman
Well-known member
Presumably a reference to Cornell's worldwide royalty-free and perpetual license as well as the right to sublicense for research use.
Personally, I am totally relaxed about that.
I took photos 25 odd years ago and was always pleased on the odd occasion when they got an audience - Naumann's Thrush (British Birds), Pallas's Grasshopper Warbler (Vinnicombe & Cottridge) and Black-headed Bunting (Isles of Scilly Bird Report cover) together with a few occasions that people bought some through a friend.
After a 20 year gap, I have started again. If anyone looks at them through eBird or Cornell use them, then I'd be equally happy. Subject to Merlin re-identifications, I believe that the attached is a Puffin (which I happily uploaded in a Farne Islands checklist) - http://ebird.org/ebird/view/checklist/S38513093 :king:
All the best
Paul
Attachments
Last edited:
Paul Chapman
Well-known member
Well I guess you hand over data (observations + rich media) as a non-monetary transaction and have access to everyone else's data (and ability to access rich media) and all the bells and whistles (organising lists, target species, trip data, Merlin etc). The data is used to do science which informs conservation.
Seems like a reasonable trade to me.
Similar threads
