/* ---- Google Analytics Code Below */

Friday, September 25, 2020

Training Machines to see 3D in the Dark

 A kind of enhanced sensory interaction.  Security an obvious application. Low light applications.

How to Train a Machine to See 3D in the Dark

Australian National University

Researchers at the Australian National University (ANU) have developed a method that uses machine learning to generate a close-to-perfect optical hologram in near-darkness. Three-dimensional (3D) holograms typically appear grainy in low light due to the shot noise limit, but ANU’s Holo-UNet was trained over thousands of learning cycles to master the appearance of an ideal hologram. Once trained, the researchers would show the Holo-UNet a hologram with missing optical information. ANU's Zhiduo Zhang said, "Much like a master painter, the machine 'remembers' how to digitally fill in those missing photons and so restore the hologram to near-perfect conditions." The researchers said their development will permit the use of holograms for purposes ranging from security to real-time imaging of living cells, with far less light than was previously required.  ... '

No comments: