Real-Time Video SuperResolution using XeSS

2 0
  • 0 Collaborators

This technology should not be limited to just gaming. This could be used as a real-time video upscaling technique and perhaps even as a compression codec. Imagine being able to have 4k or 8k video quality off of a 1080P stream. ...learn more

Project status: Concept

Artificial Intelligence

Groups
XeSS Early Access Review Group

Intel Technologies
Intel Integrated Graphics

Overview / Usage

Many times high bit-rates are required for live streaming, streaming video content, and in places like netflix, youtube, ect...

This would allow for lower overall bit-rates which would open up high quality viewing to people who have slower internet speeds around the world. This would mean that education that requires detailed video would be available to an even more broad audience then is currently allowed by technology in many areas of the world.

Methodology / Approach

I would like to use this technology at a low level to have input videos of lower resolutions and bit-rates and determine if this is viable technology to use for real-time super-resolution.

Technologies Used

I would wrap what you have in Python, as this seems to be one of the more widely used AI languages and is one that I personally am most familiar with. I will be using available hardware to me, Eluktronics Mag-15 with i7-9750h, 64gb of ram, and an RTX2070. I am also building a Desktop, as my last one died, with 128gb of ram and a 2070 super. However this still needs a few parts, motherboard and power-supply, before it will be up and running.

Comments (0)