Lots of users Google Images He must have been waiting a long time for a feature that the company announced nearly five years ago and hasn’t implemented yet: automatic colorization of black and white photos. This does not mean, however, that people are still working on it, as evidenced by some examples of images that were said to have been processed with this function some time ago.
Photo and video editing is perhaps one of the most difficult areas for developers, because it seems that today nothing works without artificial intelligence. However, this does not apply to image processing or to polishing the last pixels of detail, but above all when information that does not exist has to be created. Google Photos introduced this magic eraser feature some time ago, but they haven’t introduced another yet.
Object removal is a class in which unphotographed areas of the background must be brought in. This can only be solved by completing the advanced mode with extensive AI experience. Details are practically and theoretically impossible, because what does not exist does not exist. Surely you can reconstruct a very fuzzy picture of the Eiffel Tower – because there are millions and millions of pictures of the structure. But this is the exception and usually something that has not been photographed cannot be reconstructed.
At automatic coloring, which we brought to you a few days ago, things look a little different. Although information about colors is also not present in black and white images, one encounters completely different difficulties related to completely variable technologies. Because the coloring should result in a coherent overall picture and should of course be as realistic as possible.
The Google Photos team has been working on this feature for a number of years, and in 2019 it was included in a leaked copy of Google Photos as a beta. It had never been seen in the user realm, but was able to test the feature at the time. And to show the possibilities, our testers made it very simple: a color photo. Grayscale copy of that image and then run it through Google Images to compare the result with the original image. Check out the gallery below.
2/ Here’s a photo of my 104-year-old grandmother on her wedding day, colorized with Google Images on my phone. (You can see we have some work to do; my grandfather didn’t wear pink pants to his wedding!) pic.twitter.com/Ni8v0Bz3vg
– David Lieb (@dflieb) May 6, 2019
There are significant differences between the original image and the Google Photos image, but that’s to be expected. However, if you only see the image on the right and you don’t know which one is on the left, you will definitely be happy at first sight. It looks realistic in my opinion, it’s not too trendy nor too pastel. The result is images that are easy to look at and add even more realism to your B&W copy. Of course, one might wonder if the water below the great suspension bridge could appear red in the middle of the day without the sun setting. But red is better than brindle.
The results aren’t perfect, Google’s product manager wrote in the above tweet. But they don’t have to be like that at all, because it’s much more about heightened realism and that you can dig deeper into the situation depicted. Also, magic eraser is perfect and can never be, but the tool was released at a reasonable stage and can be further developed based on the experience of millions of users.
So I can’t understand why Google Photos has held back this feature for so long or might have stopped developing it. However, I’m sure something will come of it one day. You can find more information here.
Subscribe to GoogleWatchBlog on Google News | Subscribe to the GoogleWatchBlog newsletter
“Certified tv guru. Reader. Professional writer. Avid introvert. Extreme pop culture buff.”
More Stories
Latest technology for community safety – Celler Presse
Open day and career day on May 3: specialized lectures on renewable energies and electric mobility, modern innovative solutions in automation technology and much more.
Chassis and transmission of the G 580 with EQ