Scientists Create AI System That Can Count and Identify Wild Animals

Published By : 07 Jun 2018 | Published By : QYRESEARCH

Every day, we gain new information about various technologies and associated upgrades, mainly from the perspective of improving certain processes. And the development of artificial intelligence to automatically identify, count, and describe animals in their natural habitats is one such latest marvel in the field of technological innovations and inventions.

More Information about AI System to Identify Wild Animals

The photographs taken by the AI system are captured automatically by motion-sensor cameras, which are further be described and analyzed by deep neural networks.  The result is a system that can automate animal identification up to 99.3 percent of all images. This system also works at an accuracy of 96.6 percent when operated by crowd sourced human volunteer teams.

According to Jeff Clune, an associate professor at University of Wyoming in the United States, the AI technology helps document wildlife data in highly precise manner and without many obstacles. This is certain to transform many scientific fields such as ecology, zoology, wildlife biology, conservation biology, and animal behavior, by converting them into big data sciences. Such conversions can certainly improve the activities that involve conservation of wildlife and other precious ecosystems.

To anyone who does not know what deep neural networks are, they are mainly defined as a form of computational intelligence that is loosely inspired by how animal brains see and understand the world. These networks need large amounts of training data to work properly, and data must be accurately labelled.

The study gained relevant data from Snapshot Serengeti, which is a citizen science project that collects numerous images of animals in their natural habitats, such as lions, leopards, cheetahs, and elephants. This project was carried out in Tanzania, and numerous teams were asked to label every image manually. More than 3.2 million images were produced in this way, by more than 50,000 human volunteers, spanning over several years.

Back To Top