Google has proposed that we be able to take the best possible photos and record the best possible videos And for this, in recent years it has made available to users numerous functions such as an automatic white balance tool in Google Photos, has developed image stabilization solutions for videos and has even launched a tool that allows us to digitize photos ancient.
But now the search engine company wants to go further and is applying artificial intelligence to photography creating the form of retouch and enhance your photos before you've even taken them.
Scientists from MIT and Google are collaborating to apply machine learning algorithms in order to improve images in real time displaying them on the smartphone screen. But there is even more because it is not about automatic adjustments that are applied in the same way in all photographs, as many other applications do, but rather enhancements are tailored to individual images, according to their specific conditions.
To achieve this, the team has "trained" these neural networks with more than 5.000 images that had been retouched by five different photographers. Thanks to this Artificial Intelligence has been able to generate a formula which allows you to work on retouching photos by applying the most relevant settings for each individual image.
You can see an example in the featured image of this post, and also in the following video posted by Michael Gharbi:
This software may apply to smartphones with a minimum latency time and also with a minimal battery consumption, two factors that until now have slowed down the application of this image processing on phones.
At the moment, neither Google nor the MIT They have advanced a possible time frame for the commercialization of this technology however, it seems that it could reach Android in the not too distant future.