FFI Georgia wish to develop an application for smartphones to identify raptors that are commonly traded or poached in Georgia, eastern Europe. The aim is to provide allow law enforcement officers with an identification tool to combat illegal wildlife trade and poaching.
The app should allow a lay user to easily and accurately identify a raptor in the hand. This is different from regular bird identification apps that focus on long distance / identification in flight and that are not suitable for our purpose.
Could anybody recommend any suitable existing apps that we could work with, or are there any organisations or individuals that are developing a similar app?
If there are existing apps focusing on different regions and species that we could adjust and feed with our raptors data, that will be more efficient than us starting from scratch. Hope to learn more! Thanks in advance.
19 July 2017 10:29pm
Have you already tried the Merlin Bird ID App (http://merlin.allaboutbirds.org/)? The photo ID works pretty well. You just need to take a picture of the the whole bird. For offline mode, you can download a regional "Bird Pack", such as "US: Southeast".
Regards, Ed
20 July 2017 5:09pm
Hi Fleur,
I see now that the Merlin App bird packs are limited to North America. You could certainly contact the Cornell University Lab of Ornitology (http://www.birds.cornell.edu/) who develop the app. They may be interested in helping with such a problem.
Another option may be iNaturalist (http://www.inaturalist.org/). They have an app for logging nature observations. It is not limited to bird, but includes plants, animals, etc. They are currently adding a deep learning system to help automatically identify species from photos (currently around 10,000 species can be identified). I believe they give a top 10 possibilities. They use GPS location data from the proto or phone to provide localized suggestions. There's a demo web page here:
https://www.inaturalist.org/computer_vision_demo
According to their blog (https://www.inaturalist.org/pages/computer_vision_demo), they have integrated the computer vision module with their iPhone app. I don't think they have added it to their Android app yet. I have Android, so I have not tried the computer vision component from my app.
They may also have limited species for your area, however, the great thing about iNaturalist is that anyone can upload observations. I believe for the computer vision to be effective, they want to have at least 20 verified observations of a species by 20 different people before they add it to their model. The app is from the California Academy of Science. You may want to contact them to see if you can accelerate the process.
Best of luck!
Ed
Ed Miller
BearID Project
Arm