Those who use Pixel smartphones will know this in advance compared to others, but also those who use the GCAM coming from apk will also know what we are talking about. The next improvements that Google will introduce in the photographic field, due to the enhancement of the RawNeRF neural network, were announced during the Computer Vision and Pattern Recognition and on the tech giant's official blog. But how do night photos improve? Let's see it together.
The neural network of Google RawNerf "nerf" in the true sense of the word the photos taken in night mode. Here's how it does it in detail
According to what we learn from Google official blog RawNeRF doesn't just remove noise and improve the overall quality of a photo taken in low light conditions. This neural network is capable of much more: it can create a three-dimensional scene starting from two-dimensional images taken from different points of view. The user can also literally move the camera position to view the scene from different angles and even peek around the corner of a building.
Read also: Google changes Play Store policy to fight disinformation
Furthermore, it is possible change the AF point, focusing on any object in the scene. At the same time, the blur effect looks extremely realistic. The examples presented above in photos show that Google's RawNeRF not only removes noise, but it also restores colors and improves sharpness. It gets to the point that completely illegible text on the original image becomes quite distinguishable.
The prospects for this development and for what purposes Google intends to use it are still unclear. It is possible that one day such a function appears in Pixel smartphones and, therefore, also in imported GCams.
Through | 9to5Google
