Xiaomi presented its Xiaomi 12 a few weeks ago, a top of the range that will eventually reach European lands. One of the technologies that was not explained in the presentation is your new focus system, which they claim is inherited from CyberDog, Xiaomi’s smart dog.
Beyond its origin, the technology included in the approach of the Xiaomi 12 is interesting, and quite reminiscent of the strategies that manufacturers like Google have already carried out with their Pixel 5 and Pixel 6. In the same way, this technology it is not excluded for other members of the Xiaomi family
This is how the Xiaomi 12’s autofocus works
According to Xiaomi, one of the biggest complaints from users was the autofocus system, even on high-end models. For him, the Xiaomi 12 launches a technology that no other model had included up to date. They were already using it in the Cyberdog, under the name ‘Cyberfocus’, and they wanted to recycle it for their flagship.
Thanks to AI, the Xiaomi 12 will be able to analyze the background and the figure, temporarily storing this information. By the time you shoot, you’ll already know what to focus on
The explanation is simple, and Xiaomi now has algorithms for segment subject edges even when we are not in portrait mode. This allows that when we take a photo, the mobile can have a clearer view of the whole object on which it must focus.
In the same way, by means of artificial intelligence, they analyze the background of the image itself, to obtain complete information about the elements of the image. In this way, if, for example, we are going to photograph an animal in motion and it constantly enters and exits the composition, the Xiaomi 12 will be able to focus it quickly, because it already has figure and background information
At the moment this technology is reserved for the Xiaomi 12 family, with no news at the moment that it will reach the rest of the company’s terminals, although this cannot be ruled out. On paper, the idea is pretty good.And it is that manufacturers like Google are already using image segmentation in modes such as portrait to better focus and accurately separate the background from the figure.