Researchers at Carnegie Mellon University's (CMU) Robotics Institute have developed a computer capable of understanding body poses and movements of multiple people from video in real time, including the pose of each individual's fingers.
The team says they used a two-story dome with 500 embedded video cameras, and the insights gained from experiments in that facility can enable the identification of a pose of a group of people using one camera and a laptop.
The researchers first localized all the body parts in a scene, and then associated those elements with specific individuals.
CMU's Hanbyul Joo says the multi-camera dome was useful in not only capturing views of each individual's hands from every conceivable angle, but also for automatically annotating the hand position.
CMU professor Yaser Sheikh notes as research moves to transition from two-dimensional models of humans to three-dimensional models, this automatic annotation generation capability will become essential.
From Carnegie Mellon News
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA
No entries found