One of the greatest gifts (and curses in turn) that technology has given us is computer photography. stations like The first generation of Google Pixel showed that The phone can see more than we seeEspecially in night photography.
However, shooting in low light is quite challenging. The sensor will make too much noise If gloss is rare and here two decisions can be made: respect the noise to preserve detail, or Remove noise (wash out the image), thus losing detail. This last decision is the most common on smartphones, and Google is already training AI to solve this problem.
Google wants noise-free images, but in detail
Multinerv It is a Google project, which was launched with open source, to solve the problem of night photography. There is still a certain aversion on the part of users to digital noise, and Google’s algorithms want to fix the problem Without uploading details pictures along the way.
Google depends on NeRF (Neural Radiation Fields), a neural network developed, primarily, to generate 3D images based on 2D ensembles. The purpose of building on NeRF is, From a three-dimensional image, artificial intelligence facilitates the analysis of image information Because it can “move” through it.
“We have modified NeRF to train directly on linear raw images, while preserving the full dynamic range of the scene. By rendering the raw output images from the resulting NeRF, we can perform new High Dynamic Range Bandwidth (HDR) synthesis tasks. In addition to by changing the point of view Looking at the camera, we can manipulate focus, exposure and mapping after the fact.”
The algorithms of this model analyze the raw data of a RAW file, and use artificial intelligence to detect what information is if there is no noise in that scene. In other words, They try to preserve the details of RAW, although the noise is removed later.
Although this model is still under investigation, it is an important basis for the future of computational imaging. Let’s hope that in not too long a time, the pixels will start implementing this techniqueAnd it expands to include other manufacturers.
“Beer enthusiast. Subtly charming alcohol junkie. Wannabe internet buff. Typical pop culture lover.”