Google Research teams have developed a method that uses artificial intelligence to reduce grain in low-light photos, called RawNeRF. According to Google, this is the most powerful tool of its kind that has been created so far.
NeRF is an image synthesizer, i.e. a tool that is capable of scanning thousands of photos to reconstruct them as 3D renderings to produce a more accurate image that can be viewed from multiple angles. According to Ben Mildenhall, one of the researchers working on this project, this technology works best for photos with good brightness and a low level of grain.
Since low-light scenes proved to be problematic for this tool, people on the team had to find an alternative that would reduce this grain without reducing image quality. This is how RawNeRF was created, which with the help of artificial intelligence solves the problem of brightness. This technology can “remove” grain from an image, but without reducing its quality.
Even though it is still in the testing phase, RawNeRF could represent a particularly important innovation in the world of technology, which will give not only photographers, but all people working in the visual field the opportunity to render the world around them much more clearly .
Leave a comment