Feasibility of using convolutional neural networks for individual-identification of wild Asian elephants

Strategies of protected area use by Asian elephants in relation to motivational state and social affiliations

“…Observations were made by vehicle between 0600 to 1830 hours on tracks driven along a randomly determined route 40 . Individuals were identified through photographic cataloguing primarily using features of the ears 82 , 83 . The location of each sighting (which could include one or more individuals) was recorded using a hand-held Garmin GPS.…”

Artificial intelligence for right whale photo identification: from data science competition to worldwide collaboration

“…Neural networks are an established family of machine learning models able to learn and perform tasks by adjusting a network of artificial neurons that respond to an input layer, hidden layer(s), and then output to the next layer. Convolutional neural networks (CNN) are a variant of neural networks that have been successfully applied to computer vision problems, such as image classification (He et al 2016;Ma et al 2018;Tan et al 2019), image segmentation (Long et al 2015;Ronneberger et al 2015;Girshick et al 2018), object detection (Ren et al 2015;Redmon et al 2016;Hu et al 2017), microscopy (Xing et al 2017), and for photo identification of humans (Taigman et al 2014;Zhang et al 2017;Phillips et al 2018) and other animals (Crall et al 2013;Bogucki et al 2019;Moskvyak et al 2019;Weideman et al 2020;Cheeseman et al 2022;Clapham et al 2022;de Silva et al 2022;Blount et al 2022). As the number of layers in the neural networks increased, the term "deep learning" was coined (Dechter 1986;Aizenberg et al 2000).…”

Application of multi-species photo-ID database management systems: a user’s perspective

“…While for some animals their primary identifiable features do not change in appearance whether viewed from the left or right sides, such as the pattern of notches on the dorsal fin of many dolphin species (e.g., Würsig and Würsig 1977;Read et al 2003;Verborgh et al 2022); for many other, such as elephants where identification of individuals is based on the pattern of natural markings on their membranous ears (e.g., Ardovini et al 2008;de Silva et al 2022) or the pattern of facial wrinkles (Whitehouse and Hall-Martin 2000;Chui and Karczmarski 2022), or the coat pattern of giraffes (Bolger et al 2012;Muller 2018), or that of several species of felids (Oberosler et al 2022;Pereira et al 2022) and canids (Dorning and Harris 2019;Marneweck et al 2022), or pigmentation and scars of some poorly marked cetaceans (Karczmarski and Cockcroft 1998;Elliser et al 2022), these primary identifiable features display different patterns on the left and right sides of the animal. For that reason, DIS-COVERY offers options to verify new IDs by either a particular aspect (Fig.…”

“…There are currently numerous freely available systems for a wide range of taxa; some operate algorithms developed with a specific species in mind, while other offer a more generalised platform applicable (or adaptable) for a multi-species use (e.g., Berger-Wolf et al 2017;Brust et al 2017;Bogucki et al 2018;Maglietta et al 2020;Thompson et al 2020;Miele et al 2021;Blount et al 2022;Clapham et al 2020Clapham et al , 2022. Some have proven to be remarkably successful, facilitating automated comparisons of large photo-ID datasets (hundreds of thousands individual IDs) across large geographic scales (e.g., Cheeseman et al 2022), while other remain at present impractical (e.g., de Silva et al 2022) or insufficiently reliable (e.g., Morrison et al 2016) for field research applications. Given the developmental changes in feature appearance and mark changes over time (e.g., Carlson and Mayo 1990;Waye 2013;Wattegedera et al 2022), the concepts of active learning with human-in-the-loop and continuous lifelong learning (Käding et al 2016;Bodesheim et al 2022) come especially handy.…”