I remember reading about a revolutionary algorithm that can make something out of nothing. Traditional MRI scan is not efficient in some cases, and sometimes produces a lo-res and grainy photo that drastically drops data accuracy. But according to Wired, we now have a clearer visual sense of what's on paper and what else's to come...
"By using mathematical concept called 'sparsity', the compressed-sensing algorithm can take a fuzzy, lo-res file and transform it into an image that is sharp and crisp" - Wired, 18.03
How it works is that the algorithm will modify the picture in stages by laying coloured shapes over the randomly selected image. It intelligently detects the adjacent pixels in the same area and fill in the same colour and does this over and over again with smaller and smaller shapes to finally achieve a crisp clear image that is near-perfect.
I can't help but wonder if there is a similar magic working here, as shown in this video of Adobe Photoshop's Content-Aware Fill tool. You select the area which you want to clean, and it detects the possibility of adjacent hues in ultra small fine tiny little wee bits and fill in what it 'thinks' could be logical. Five seconds of dragging your mouse around, one click, and viola. New picture. Like magic ;)
Anyone know if the same technology is employed here?
Subscribe to:
Post Comments (Atom)
Posts
No comments:
Post a Comment