A clear example can be found in the difficulties presented by low-quality
Posted: Wed Jan 08, 2025 3:55 am
We live in the digital age . A time that has allowed us to make great advances that until recently seemed like science fiction. In a world as visual as ours, these steps we have taken have been of great help, but the truth is that there is still much to do.
images that we can search for on the Internet.
Surely you have seen in more than one series or movie how investigators, with a simple zoom gesture, have been able to solve who is behind the blurry pixels (something difficult to achieve in real life). Now Google has the solution.
As reported by El Confidencial , the search saudi arabia phone data giant has achieved this by using Google Brain , an artificial intelligence software with which the company is able to convert small photographs that tend to appear pixelated into more detailed images. All by using two neural networks.
The system used starts with a low-resolution image and reduces other good-quality images to its size to determine which ones are most similar. From this point, the PixelCNN conditional image generation system is used, capable of adding much more realistic details.
The result achieved by combining both neural networks offers us an image of higher quality than the original . We must make it clear that the resulting image is not the original, but that the system offers us a new one, “completing” what it lacks. Of course, we are facing one of those great and necessary advances that we talked about at the beginning of this news.
images that we can search for on the Internet.
Surely you have seen in more than one series or movie how investigators, with a simple zoom gesture, have been able to solve who is behind the blurry pixels (something difficult to achieve in real life). Now Google has the solution.
As reported by El Confidencial , the search saudi arabia phone data giant has achieved this by using Google Brain , an artificial intelligence software with which the company is able to convert small photographs that tend to appear pixelated into more detailed images. All by using two neural networks.
The system used starts with a low-resolution image and reduces other good-quality images to its size to determine which ones are most similar. From this point, the PixelCNN conditional image generation system is used, capable of adding much more realistic details.
The result achieved by combining both neural networks offers us an image of higher quality than the original . We must make it clear that the resulting image is not the original, but that the system offers us a new one, “completing” what it lacks. Of course, we are facing one of those great and necessary advances that we talked about at the beginning of this news.