The Bird Image Identification System. A useful tool for wildlife photographers, bird/nature enthusiasts, researchers and academics.
Upload your bird images below, and the system will try to recognition them using Deep Learning/AI and Computer Vision.
Do you find this site useful? Let us know your feedback on how we can improve it.
Changelog: v0.3.3 8-Jul-2019
- Added support for 10 new bird species from 0.3.2: Black Drongo, Black Eagle, Black Kite, Black Noddy, Black Stork, Black Tern, Black-and-Orange Flycatcher, Black-crested Bulbul, Black-crowned Night Heron, Black-headed Gull, bringing up the total to 274 species of birds. See the About page for the full list.
- Tweaked the convolutional neural network model and changed the training technique for better accuracy.
- Added support for 35 new bird species: Alpine Swift, Amur Falcon, Ashy Drongo, Ashy Prinia, Ashy Woodswallow, Ashy-crowned Sparrow Lark, Asian Brown Flycatcher, Asian Emerald Cuckoo, Asian Fairy Bluebird, Asian Koel, Asian Openbill, Asian Palm Swift, Asian Paradise-flycatcher, Asian Pied Starling, Asiatic Dowitcher, Banded Bay Cuckoo, Bar-tailed Godwit, Bar-winged Flycatcher-shrike, Barn Owl, Barn Swallow, Bay-backed Shrike, Baya Weaver, Besra, Black Bittern, Black Bulbul, Black-capped Kingfisher, Brahminy Kite, Brown Fish Owl, Ceylon Paradise-flycatcher, Chinese Pond Heron, Great Egret, Grey Heron, Intermediate Egret, Little Egret, Orange-breasted Green-pigeon.
- Added a new experimental update to the neural model.
Info for the users:
- The system uses a deep learning model (specifically, a convolutional neural network), which has been trained on thousands of known images, to classify the images you upload.
- As this is a beta release, the system is still limited. Therefore, miss-identification of images are still possible.
- Over time, as we train the system further, the accuracy will get better, and the miss-classifications will reduce.
- The system can currently identify 274 species of birds. See the About page for details.
- We will not store any of the images you upload. The images are discarded as soon as the classification is made.
- The code of this application is publicly available at GitHub. Feel free to go through the code, and suggest improvements.