Re:New algorithm unlocks high-resolution insights for comput...
by 3abiton on Tuesday March 19, 2024 @07:22PM
(#39761350)
Very interesting, I am curious how do people reach that train of thought to a successful idea. So many great algorithms based on small twists.
Re:New algorithm unlocks high-resolution insights for comput...
by kreelman on Tuesday March 19, 2024 @07:45PM
(#39761518)
It is interesting indeed.
One wonders if the researchers of this particular bit of work made it mandatory to go for walks at lunch and think about how their vision chunked/filtered the information it was receiving.
Interesting that they "perturb" the image to get some noise involved. I'll need to read it over again.
Re:New algorithm unlocks high-resolution insights for comput...
by davecanderson on Tuesday March 19, 2024 @10:23PM
(#39762358)
Nature is such a good source of inspiration, the "perturb" approach reminded me of [fixational eye movement][1] but maybe that's only a clear link in retrospect.
Re:New algorithm unlocks high-resolution insights for comput...
by cryptonector on Tuesday March 19, 2024 @10:56PM
(#39762504)
This seems like it could have been inspired by how human vision works.
Re:New algorithm unlocks high-resolution insights for comput...
by Grieverheart on Wednesday March 20, 2024 @04:42AM
(#39764043)
Also similar to temporal antialiasing https://en.wikipedia.org/wiki/Temporal_anti-aliasing .
Re:New algorithm unlocks high-resolution insights for comput...
by jeffdaniels27x on Wednesday March 20, 2024 @09:08AM
(#39765900)
Perhaps similar in some ways to how big cats' eyes reflect the light back from behind the retina (i.e. back through it for a second pass) to capture more light. I'm sure I heard that on a nature documentary ...
They reproject the input images and run the low-res network multiple times. Then they use an approach similar to NeRF to merge the knowledge from those reprojected images into a super-resolution result.
So in a way, this is quite similar to how modern Pixel phones can take a burst of frames and merge them into a final image that has a higher resolution than the sensor. Except that they run useful AI processing in between and then do the super-resolution merge on the results.