Fast method for noise level estimation and integrated noise reduction
Angelo Bosco, Arcangelo Ranieri Bruna, Giuseppe Messina, Giuseppe Spampinato · IEEE Transactions on Consumer Electronics · 2005
This paper describes a fast method for noise level estimation and denoising. Specifically, we address the problem of estimating the standard deviation of additive white Gaussian noise in digital images; the computed value is used to reduce Gaussian noise and eliminate defective pixels in a raw digital image. The method is particularly suitable for implementation in low power mobile devices with imaging capabilities such as camera phones, as well as digital still cameras (DSC).