wesm Loc: Los Altos CA

Even though it's Wired, it's a pretty good article. It's about training a machine to look at what others have done to post-process photos. The cool part, at least for a techie, is that it can happen in 1/20th of a second, so your smartphone can actually show you in the preview before you press the shutter button.

I saw a demo of something similar 5 years ago at a tech talk. Guy had a picture of a lake, but some rooftops obscuring the near shore. He ran an algorithm against some 100,000 photos of lakes and shorelines, and came up with a composite that removed the rooftops, and looked very believable.

No, it won't replace the creative process (at least, not yet). But it does beg the question, what is a photo? I suppose it's just a logical extension of the discussions here about what SOOC really means, or whether analog and digital are the same

https://www.wired.com/story/googles-new-algorithm-perfects-photos-before-you-even-take-them/

I look forward to thoughtful responses.

Leicaflex Loc: Cymru

Interesting article.
Thank you for the link.

cthahn

This is what many hotographers would like to have. Push a button and have a perfect picture.

machia Loc: NJ

Gimmick for non-photographers basically . That's how I see it anyway .

Martino Loc: Northwest Florida

Sorry, what is SOOC?

A photographer is anyone who uses a photographic process to create an image. Nothing more.

Some people produce better images than others. Some create technically good images. Some create works of art that bear varying degrees of difference from that simply captured by the photographic or image capturing processing.

TonyBot

MIT Media Lab is light years ahead of what is really happening in the real world - but most of it's findings have had a very high level of occurring, sometimes decades into the future. Neural networks, back in the ancient world (you know, the 1980s, when the UU net was just beginning) could solve the "10 city" problem more quickly than any probably any computer in the world. The "100 city" problem was most likely solved before the millennium by neural networks. (briefly, the "x city" problem is designed to find the quickest and shortest route between "x" number of points) So ... I wouldn't worry about a mostly biological computer taking over photography ... yet. ;)