I'm pretty sure that if focus could be corrected through software alone, that feature would have been added to every picture-taking device in the world.
Just noticed that comment scanning an old post, and NVidia just recently showed this exact thing off. It does require a special multi lens camera, and some heavy GPU lifting currently. Still a ways off from consumer level tech, but a fun glimpse at the possible future for photography.
Skip to 2:50 if you just want to see the demo instead of the discussion before it.
Just noticed that comment scanning an old post, and NVidia just recently showed this exact thing off. It does require a special multi lens camera, and some heavy GPU lifting currently. Still a ways off from consumer level tech, but a fun glimpse at the possible future for photography.
This isn't a new idea. Wavy fronted lenses combined with massive amounts of processing have been around for years. Slashdot had a story about it but the company got bought by Omnivision and I can't find the gallery. It was similar to what is shown in your clip though.
I think the new part here (and the reason it was at an NVidia conference) is that the computing power needed is now at a level to be able to do it real time, with a common GPU.