Researcher chirpy about bird-app prospects

University of Otago doctoral candidate Tapabrata Chakraborti has undertaken research that could...
University of Otago doctoral candidate Tapabrata Chakraborti has undertaken research that could lead to the development of a smartphone app to identify New Zealand birds. Photo: Christine O'Connor
University of Otago doctoral candidate Tapabrata Chakraborti is excited that his research is paving the way for a smartphone app enabling tourists to identify "our national treasure" of birds.

Indian-born Mr Chakraborti recently completed his computer science doctorate and his research findings had been "well received in leading peer-reviewed international conferences", he said.

New Zealand was home to many endemic or native birds, and many of them were also threatened or endangered.

These birds were of interest not only to experts, but also to bird-watchers  and the general public.

Tablets and mobile phones now made it potentially possible for anyone to become a bird-watcher.

But this had required research into adopting artificial intelligence and machine learning techniques and developing robust apps.

The Otago computer science department had recently been conducting research in this field, mainly through Mr Chakraborti’s doctoral work and his supervisors, Drs Brendan McCane and Steven Mills.

The main challenge was to "make the algorithm learn the subtle differences between similar-looking birds" such as kea and kaka.

The project’s "future and final goal" would be to translate these algorithms into a user-friendly app that could operate as a digital bird-watcher.

A user would photograph a bird on a mobile device and the app would suggest the bird’s name.

New Zealand research could also be integrated into an existing bird app, such as one operated by Cornell University, in the United States, which mainly focused on North American and European birds.

Mr Chakraborti would also, from March, undertake postdoctoral study at Oxford University, England, on the use of artificial intelligence in medical imaging.

He hoped other researchers would further develop his bird-recognition work.

Add a Comment