Researchers use deep transfer learning to study nest site fidelity in painted stork in Delhi zoo

A squad of researchers person successfully utilized heavy transportation learning (DTL), a non-invasive artificial intelligence-based tool, to survey nest tract fidelity successful painted stork (Mycteria leucocephala) successful the National Zoological Park, Delhi.

The researchers chose a antheral painted stork with a distinctive wounded scar connected its cervix and observed it implicit 4 consecutive breeding seasons from 2022 to 2025 to measure the nest tract fidelity – a trait of utilizing the aforesaid nesting tract successful successive breeding seasons.

They named the stork ‘Ringo’, successful honour of drummer Ringo Starr of iconic popular euphony set The Beatles, for the study. A full of 2,349 precocious solution images of Ringo, covering some sides and helping markings successful the folded position, were photographed for the study. They besides clicked 1,755 images, showing some the near and close sides of the wings of different nesting storks. These images collected during breeding seasons from 2022 to 2025 were utilized arsenic a instauration for studying idiosyncratic recognition done unsocial features contiguous successful painted storks.

As per the study, published successful Royal Society Open Science, the researchers employed a non-invasive attack to show Ringo. They utilized scale-invariant diagnostic alteration (SIFT), a machine imaginativeness algorithm that extracts distinctive features from images for close matching and idiosyncratic identification. The SIFT features identified the unsocial scar marking successful Ringo images.

The researchers besides developed a DTL model, to place features that separate Ringo from different storks, with the feather signifier serving arsenic a signifier of biologic fingerprint. The instrumentality validated Ringo’s individuality with 98% accuracy and the bird’s repeated sightings astatine the aforesaid spot implicit 4 consecutive years confirmed its nest-site fidelity.

The survey was conducted by a squad comprising Abdul Jamil Urfi and Paritosh Ahmed from the Department of Environmental Studies, University of Delhi; Mylswamy Mahendiran from the Division of Wetland Ecology, Salim Ali Centre for Ornithology and Natural History, Coimbatore, and Mylswamy Parthiban from Department of Physical Sciences and Information Technology, Agricultural Engineering College and Research Institute, Tamil Nadu Agricultural University, Coimbatore.

As per the study, the findings item the imaginable of pattern-based designation and usage of DTL arsenic a almighty and non-intrusive instrumentality for semipermanent monitoring of assemblage waterbirds.