LEAP Tracking Giraffe Motion through Deep Neural Networks (VIDEO)

Princeton University

Caption

Princeton researchers created LEAP, a flexible motion-capture tool that can be trained in a matter of minutes to track body parts over millions of frames of video with high accuracy, without any physical markers or labels. Here, graduate student Talmo Pereira took giraffe footage from the Mpala Research Centre's live video feed, labeled 30 frames to train LEAP's neural network, and then LEAP generated this within seconds.

Credit

Left: Raw video footage courtesy of mpalalive.org Center and right: Courtesy of the researchers

Usage Restrictions

Editorial use only

License

Licensed content

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.