We all know someone that has an older photo taken in black and white and many of us have seen those request to transform those photos into color. Converting old black and white photos to color may have just gotten easier with a new app.
Very talented Photoshop artists can spend many hours manually converting black and white photos into color, or you can try some of the other software out there with little success. Now there is a new app called the Interactive Deep Colorization, developed by Richard Zhang and the rest of his team from the University of California at Berkeley.
How does this app differ from the rest? Well it can make its predictions on what colors should be used, from there is where the magic starts. You can make changes to its predictions, customizing the color using a suggested color palette or choosing your specific colors for areas throughout the photo. The software’s AI has studied millions of images for its predictions and how the colors can be applied. It may not get the colors right at first with the exact colors you wish to use, but now it’s much easier to add color to your favorite black and white photos. This will surely reduce Photoshop users' seat time behind the screen adding color and maybe this will be picked up in the future and added to Photoshop.
This is an odd sentence "It may not get it right at first, or the exact colors you wish to use, but not it’s much easier to add color to your favorite black and white photos. " . That sentence aside, I really enjoyed the article and I can't wait to try this tool out for myself.
Thanks, but of miss types that and didn't catch it before.
Do you proofread anything?
The concept looks interesting. What is missing is simplicity. They could not make it more difficult to get the app downloaded. For most, unless you are a developer, "using GIT" or "checking out with SVN" does not mean anything. What do they mean by "Deep learning software"? A long learning curve or software that does deep analysis? Why not make it simple rather than making people guess?