Saturday, January 14, 2017

Google RAISR and Draco Reduce Amount of 2D and 3D Data

Google says it has developed RAISR machine learning technology that can upsample the image by 4x with visually pleasing results:

Top: Original, Bottom: RAISR super-resolved 2x.
Original image from Andrzej Dragan

The smart machine learning algorithm is said to be able to recognize and remove the aliasing artifacts from the downsampled image:

Left: Low res original, with strong aliasing.
Right: RAISR output, removing aliasing.

Google paper "RAISR: Rapid and Accurate Image Super Resolution" by Yaniv Romano, John Isidoro, Peyman Milanfar is to be published in IEEE Transactions on Computational Imaging.

Google also announces Draco - an open-source algorithm for 3D image compression.

"With Draco, applications using 3D graphics can be significantly smaller without compromising visual fidelity. For users this means apps can now be downloaded faster, 3D graphics in the browser can load quicker, and VR and AR scenes can now be transmitted with a fraction of the bandwidth, rendered quickly and look fantastic."

No comments:

Post a Comment

All comments are moderated to avoid spam.