Auto-ID of birds in photos—now for UK and Ireland (1 Viewer)

[idaviesbird]

Merlin Bird ID is now available for the first batch of birds in Europe. Answer a few easy questions or snap a photo, and get real-time answers to your bird ID questions. Europe: Great Britain and Ireland features 250 species with ID text written by Steve N.G. Howell, stunning images and audio from your eBird contributions that are archived in the Macaulay Library, and Merlin’s well-known Photo ID. Stay tuned in the coming months for coverage of more uncommon species in the UK and Ireland as well as more of Europe. Give Merlin a try today.

While Merlin is primarily known for its identification features, there is even more great functionality that cements Merlin as an invaluable addition to any birder’s phone. Regional bird packs give you instant access to songs and calls for each species you might encounter, hand-picked by the Macaulay Library archivist team. Additionally, Merlin provides quick access to eBird data in the form of custom ‘likely lists’—just choose your location anywhere in the world and get a list of the most likely birds to encounter in the area.

Photo ID now covers 1,700 North American and European birds which includes all the species that will be included in two additional packs covering Northern and Western Europe. These packs will be out in time for the British Birdfair, so if you happen to be there, stop in at the Cornell Lab of Ornithology booth, grab some Merlin and eBird stickers, and say hi!

 

[Surreybirder]

'Like' as they say nowadays.

 

[fugl]

Christ, takes all the sport out of it. . ..

 

[Surreybirder]

Not sure I agree. My guess is that the software becomes pretty unreliable for the 'difficult' species. Male red-backed shrike is pretty much impossible to get wrong but what about a worn female?

 

[fugl]

Not sure I agree. My guess is that the software becomes pretty unreliable for the 'difficult' species. Male red-backed shrike is pretty much impossible to get wrong but what about a worn female?

Unreliable for the moment, sure, but for how long?

 

[Farnboro John]

Unreliable for the moment, sure, but for how long?

As long as birds can get wet, have displaced feathers, be in deep moult, not be the common species, be at a funny angle, be photographed incompetently......

I'm sure I've missed many scenarios. My concern is that the software will likely give an answer regardless, and the less experienced will be inclined to trust it rather than the experienced birder telling them different....

My other concern (this may go on all night) is that just as people used to know phone numbers but now don't because they select the menu item that says "Dave D mob", or now can't navigate or contradict a satnav, this will actually diminish field ID skills overall as the beginner relies on s/w instead of learning stuff.

John

 

[Paul Chapman]

My concern is that the software will likely give an answer regardless

John

Like the Bird Identification Forum.

Happy Birding

 

[fugl]

Like the Bird Identification Forum.

Happy Birding

Bingo!

 

[fugl]

As long as birds can get wet, have displaced feathers, be in deep moult, not be the common species, be at a funny angle, be photographed incompetently......

Well, we'll see. With the tremendous (and ever increasing) computing power of mobile devices, i can easily envisage the day when ID apps outperform "expert" birders as easily as computers best chess grand masters today. Just a matter of database access and comparing this with that, after all, things that computers are supremely good at.

In fact, I can even see coming to like it after a while; pointing my phone at an animal or plant (or any other real-world object for that matter) not just for the "ID" but for anything else I might be interested in knowing about it.

 

[Nutcracker]

Depends of course on the reliability of the "stunning images and audio from your eBird contributions that are archived in the Macaulay Library" - if any of those are misidentified, you may get wrong answers.

 

[Farnboro John]

Well, we'll see. With the tremendous (and ever increasing) computing power of mobile devices, i can easily envisage the day when ID apps outperform "expert" birders as easily as computers best chess grand masters today. Just a matter of database access and comparing this with that, after all, things that computers are supremely good at.

In fact, I can even see coming to like it after a while; pointing my phone at an animal or plant (or any other real-world object for that matter) not just for the "ID" but for anything else I might be interested in knowing about it.

I entirely agree with the second half of your post: after all I have a mobile phone and use a satnav regularly, and I would certainly interrogate this app for species I am less familiar with.

The thing is, I do have knowledge about not only the birds and mammals that are central to my hobby but also about animal classification and identification in general. Just as my taught ability to use maps informs me to be able to critically assess satnav information, I am able to take a considered view of ID info even where I'm not at ease with the species group: this app will have the Internet peppered with misidentified rubbish photos clouding recorders' and researchers' jobs before you can say Bogota Sunangel.

John

 

[Paul Chapman]

Depends of course on the reliability of the "stunning images and audio from your eBird contributions that are archived in the Macaulay Library" - if any of those are misidentified, you may get wrong answers.

Of course, the identification will presumably work from the totality of the photos to construct an analysis so the 9,462 photos of House Sparrow rather than an odd one that may be misidentified:-

https://ebird.org/media/catalog?dat...cus&species=House Sparrow - Passer domesticus

Of course, I have used this as an extreme example. There are 17 labelled as Basra Reed Warbler:-

https://ebird.org/media/catalog?dat...s=Basra Reed-Warbler - Acrocephalus griseldis

Just as my taught ability to use maps informs me to be able to critically assess satnav information, I am able to take a considered view of ID info even where I'm not at ease with the species group: this app will have the Internet peppered with misidentified rubbish photos clouding recorders' and researchers' jobs before you can say Bogota Sunangel.

John

It is common to see trends within other taxa of misidentifications between different species pairs such as a raft of Red-green Carpets called Autumn Green Carpets. This often happens within locations so exposure of an overall data set of photos would actually help to reduce errors.

I do wonder though whether any of negative comments have been based on trying the software which will presumably learn from errors or misidentified photos. I need to revisit some of my Isabelline/Red-tailed Shrikes but most of mine I'll relatively happy with:-

https://ebird.org/media/catalog?dat...date_desc&userId=USER674479&user=Paul Chapman

All the best

 

[Nutcracker]

Of course, the identification will presumably work from the totality of the photos to construct an analysis so the 9,462 photos of House Sparrow rather than an odd one that may be misidentified:-

https://ebird.org/media/catalog?dat...cus&species=House Sparrow - Passer domesticus

Of course, I have used this as an extreme example. There are 17 labelled as Basra Reed Warbler:-

https://ebird.org/media/catalog?dat...s=Basra Reed-Warbler - Acrocephalus griseldis

Obviously House Sparrow isn't likely to be a problem, but if suppose, 15 of those 17 Basra Reed Warbler pics come from the people cited in this thread, there could be very real problems . . .

 

[McMadd]

Obviously House Sparrow isn't likely to be a problem, but if suppose, 15 of those 17 Basra Reed Warbler pics come from the people cited in this thread, there could be very real problems . . .

If yer going for Basra RW I doubt you'll be carrying one of these things...

 

[Andy Adcock]

Like the Bird Identification Forum.

Happy Birding

Love it


A

 

[a_j_steele]

One thing searching for the app on the playstore does't make it clear its uk birds as well as america birds

 

[Surreybirder]

Of course, the identification will presumably work from the totality of the photos to construct an analysis so the 9,462 photos of House Sparrow rather than an odd one that may be misidentified:-

https://ebird.org/media/catalog?dat...cus&species=House Sparrow - Passer domesticus

Of course, I have used this as an extreme example. There are 17 labelled as Basra Reed Warbler:-

https://ebird.org/media/catalog?dat...s=Basra Reed-Warbler - Acrocephalus griseldis



It is common to see trends within other taxa of misidentifications between different species pairs such as a raft of Red-green Carpets called Autumn Green Carpets. This often happens within locations so exposure of an overall data set of photos would actually help to reduce errors.

I do wonder though whether any of negative comments have been based on trying the software which will presumably learn from errors or misidentified photos. I need to revisit some of my Isabelline/Red-tailed Shrikes but most of mine I'll relatively happy with:-

https://ebird.org/media/catalog?dat...date_desc&userId=USER674479&user=Paul Chapman

All the best

I wouldn't presume to question your shrike ID's, Paul, but you have some lovely photos there.

 

[ZanderII]

Not sure I agree. My guess is that the software becomes pretty unreliable for the 'difficult' species.

Why guess? You could download and test it... it does semi-p vs western, AGP vs PGP better than most birders...

 

[ZanderII]

Obviously House Sparrow isn't likely to be a problem, but if suppose, 15 of those 17 Basra Reed Warbler pics come from the people cited in this thread, there could be very real problems . . .

So you are saying those aren't Basra RW? Are these the authors of the fabricated data papers?

This seems like pretty lazy criticism all round.... why not have a go at testing Merlin... and re-identifying the Acros if need be.

 

[Nutcracker]

So you are saying those aren't Basra RW? Are these the authors of the fabricated data papers?

No, I was postulating 'what if they were' :t: