Real Noise Reduction

About

Description: Often images taken with smartphones or point-and-shoot digital cameras come out noisy due to lack of sufficient lighting. This low-light noise problem is widespread, being present in all of the smartphones in the world, more than 1 billion total. This problem generates in consumers disappointment and frustration with the quality of the images taken in low light. While a number of commercial denoising packages are already available on the market, the majority of them are trained on images corrupted by artificial noise, rather than trained on real low-light noisy images. Since artificial noise has different characteristics than real noise, these packages do not perform as well as a denoising algorithm trained on images corrupted by real noise as we have created. We have developed a fully automatic state of the art algorithm (RENOIR) for denoising smartphone and digital camera images which have low-light noise problems. The RENOIR algorithm could be either; be sold directly to the public as a standalone application or could be licensed to smartphone or digital camera manufacturers to be embedded in their devices.