Anil Jain and colleagues at Michigan State University have developed a mobile application that can identify and track primates in the wild, identifying individual animals from facial photos.
The team captured about 3,000 photos of the primates at North Carolina's Duke Lemur Center, in addition to thousands of images of golden monkeys and chimpanzees provided by conservationists.
A neural network learned to differentiate the primates' facial features, and proved especially effective at identifying lemurs; it achieved more than 80% accuracy even when it had to infer a specific animal could not be recognized because it was not included in the training dataset.
Jain says the researchers also determined the accuracy for recognizing chimpanzees was lower, despite a larger training dataset, due to the poorer quality of those images.
From New Scientist
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found