Google Research is working on a new photo-based noise control tool based on artificial intelligence algorithms. It is expected that the new solution will allow the qualitative reconstruction of dark scenes, leaving only minimal artifacts, and it will be much more effective than the current solutions.
The elements of artificial intelligence in photo processing have been in use for a long time, especially in smartphones and software for photo processing after filming. The possibility of high-quality photo noise elimination is one of the most needed functions, as even the best sensors cannot be removed completely without artifacts, especially if there is a lack of lighting.
The technology presented by Google Research allows photographers to "see in the dark." The new RawNeRF tool is part of an open-source project with MultiNeRF. It will be particularly useful for those who start shooting dark scenes — it will be possible to do so with a minimal loss of quality.
NeRF technology is a synthesizer capable of scanning multiple RAW images and retrieving an accurate 3D brand. According to Google data, RawNeRF technology is able to combine images from different angles to eliminate noise and reconstruct the scene. The display, display, and focus can change.
A smartphone photo is used to demonstrate the capabilities of RawNeRF. It is noteworthy that I.I. algorithms are used for processing that are "trained" not on the JPEG already processed but on the RAW content.
The advent of new technology in smartphones and cameras can indeed change the rules of the game, both among professional photographers and among amateurs with smartphones. Techspot states that as long as there are no noise control tools in the market, they can ensure comparable image quality.