Facial recognition software is gradually becoming commonplace online, especially given the advances in AI, but why should humans be the only primates that take advantage of it?
A team of lemur biologists and computer scientists has created LemurFaceID – recognition technology that can identify facial variations in lemurs with 97 per cent accuracy.
Identifying individual lemurs has previously relied upon cataloguing unique identifiers, such as differences in body size and shape, or the presence of injuries and scars. This method is not entirely reliable, though, as zoologists are forced to rely on factors that can change over time, or are perhaps too common amongst the localised lemur population. This imposes limitations upon long-term, multi-generation studies of lemur populations.
Using a dataset of 462 images of 80 red-bellied lemurs, and a database containing a further 190 images of other lemur species, the inter-disciplinary team of scientists was able to modify human facial recognition software. Many lemur faces possess unique features such as hair and skin patterns that computer systems can be trained to recognise - much like individual features found in human faces.
Study author Doctor Rachel Jacobs, from George Washington University, said the team was "surprised with the high degree of accuracy" and a review in the BMC Zoology journal found the software could detect individual faces with an accuracy of 97 per cent.
LemurFaceID could allow for the animals to be recorded and monitored with minimal interference from biologists, which in turn, could reduce costs. While initial uses of the software have taken place in captive settings in Ranomafana National Park, Madagascar, the promise of using facial recognition in the wild could be of great benefit to the global lemur population.
Monitoring lemurs in the wild is crucial to developing effective conservation efforts. Out of the 50 varieties of lemurs, 10 are critically endangered, seven are endangered and 19 are considered vulnerable. Habitat loss is a significant contributor to the reduction in lemur numbers - therefore, studying lemurs in their natural environment is critical to understanding how they cope with habitat changes, as well as developing methods to help protect the lemur population.
The team behind LemurFaceID believes the software could also be used for identifying other primate and non-primate species with variable facial hair and skin patterns, such as bears, red pandas or sloths. A larger dataset of individuals and photographs is needed to boost the accuracy of identification in non-captive settings, but the first variation of the software shows promise for further cross-species identification in the future.
This article was originally published by WIRED UK
