Logo Logo
deutsch
directory schraeg
Knowledge
Codecs
Hardware
Camcorders· Cinema-Kamera· Computers· Drohnen· GPU· Kamera-Zubehör· Video-DSLR· accessories
Software
Compositing· Color correction· DV Editing
DV-Movies

HowTo
Shooting· Sound
Misc
Forschung· Reviews· Streaming
/// News
Google'amp;s new super-resolution algorithm SR3 upscales faces almost perfectly

Google'amp;s new super-resolution algorithm SR3 upscales faces almost perfectly

[18:11 wed,14.July 2021   by ]    

A team of Google researchers has unveiled a new DeepLearning super-resolution algorithm that puts all previously developed methods to shame. Dubbed SR3 (Super-Resolution via Repeated Refinement), the method demonstrates its power especially in examples of upscaled faces: photos of faces with a resolution of only 64 x 64 pixels are upscaled in two steps, first to 256 x 256 pixels, then to 1,024 x 1,024 pixels, which corresponds to a 16x magnification. In another experiment, images of objects such as flowers, fire engines, birds, or buildings are scaled up from 64 × 64 to 256 × 256 pixels.

SR3-Example2




The high quality of the upscaled images was demonstrated in an experiment: test subjects were asked to decide in a blind test whether the high-resolution original image or the image that was first reduced in resolution and then upscaled again using SR3 looked better - in an 8-fold upscaling from 8 × 8 to 128 × 128 pixels, about half chose the upscaled face.

Cascade
Upscaling in several steps


This corresponds exactly to the random distribution and thus says that it is no longer possible to distinguish between the original and the version generated by Super Resolution. In the much more difficult task of upscaling a 64 x 64 photo of a natural object 4 times to 256 x 256, 40% of the test participants still preferred the generated image to the original.



How does Super Resolution even work?



Until the last decade, the dogma was that when resolution is increased, no detail can be added that is not in the original image, but for the last few years we have been in a paradigm shift. The buzzword is super-resolution. Super-resolution refers to technologies that add more detail to an image in post-processing than is originally available in digital form.

Simply put, they are processes that increase the resolution of an image, using more than just interpolation between known pixels for the "extra pixels." Thus, a super-resolution (SR) application must coherently add details.

SR3-Example1
More upscaled objects in comparison.


With the advent of Deep Learning, Super-Resolution has received a real boost as AI algorithms become very good at recognizing and adding to objects. For example, if an AI has seen millions of faces from different angles and in diverse lighting situations, it can subsequently add learned details to each face in an image. However, this is just as true for plants, cars or animals.

Neural networks are trained with many images and thus recognize typical natural structures/patterns - be it those of faces or other objects. When upscaling, this regularity is used - i.e. structures in the low-resolution images are reconstructed in a meaningful way when upscaling. More about this topic in our article More image sharpness by artificial intelligence - Super-Resolution.

super_res_examples


Adobe has such a super-resolution algorithm already integrated in Photoshop and the paid tool Video Enhance AI as well as TecoGAN do something similar for videos.

The new SR3 algorithm ( here the corresponding research paper) applies a special trick and upscales the images in several passes which leads to better results than interpolating from the base image to the high-resolution version. Or more specifically in the researchers& words, "SR3 adapts denoising diffusion probabilistic models to conditional image generation and performs super-resolution through a stochastic iterative denoising process."

Link more infos at bei iterative-refinement.github.io

deutsche Version dieser Seite: Googles neuer Super-Resolution Algorithmus SR3 skaliert Gesichter nahezu perfekt hoch

  

  Vorige News lesen Nächste News lesen 
bildPanasonic GH5 II - gleicher Sensor, bessere Dynamik? bildYouTube Shorts macht TikTok Konkurrenz - Aber Achtung YouTube Creators!


related news:1E0New firmware turns RØDE Wireless GO II transmitter into field recorder 1.April 2021
New Dell UltraSharp U4021QW: 40-inch monitor with 5K resolution and Thunderbolt 3 6.January 2021
Zoom F2 Field Recorder - extra small with 32-bit float technology 11.November 2020
Neural network DAIN interpolates stop-motion frames 9.September 2020
TecoGAN: New AI Super-Resolution Algorithm makes video 4x sharper 6.September 2020
Canon EOS R5 and R6 - Stabilized full frame up to 8K RAW with Dual Pixel AF 9.July 2020
8K and Dual Pixel AF in new Samsung ISOCELL GN1 smartphone camera sensor 25.May 2020
alle Newsmeldungen zum Thema Basics
1E0Flim: AI-based film still search gets intelligence upgrade 4.June 2021
Vivo X60 Pro 5G Smartphone -- gimbal-stabilized camera with Zeiss lens 21.May 2021
Hollywood composer Hans Zimmer designs Fahrgeräusche für BMW 13.May 2021
Hype alarm: sell your film as NFT with blockchain tech and get rich fast? 9.May 2021
Oscars 2021 - more modern and diverse: Nomadsland wins Best Film and more 26.April 2021
Sony FE 14 Millimeter F1.8 GM -- compact ultra wide angle lens announced 22.April 2021
RØDE Connect: Free podcast and streaming tool for up to four NT-USB mini mics 17.April 2021
alle Newsmeldungen zum Thema Film


[nach oben]
















Archiv Newsmeldungen

2021

July - June - May - April - March - February - January

2020
December - November - October - September - August - July - June - May - April - March - February - January

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000






































deutsche Version dieser Seite: Googles neuer Super-Resolution Algorithmus SR3 skaliert Gesichter nahezu perfekt hoch



last update : 28.Juli 2021 - 15:36 - slashCAM is a project by channelunit GmbH- mail : slashcam@--antispam:7465--slashcam.de - deutsche Version