Nonlocal mean image denoising method based on Poisson distribution

In order to handle the image denoising problem, a non-local mean image denoising method based on Poisson distribution is proposed in this paper. Each pixel in image is modeled as a poisson distribution. The parameters of Poisson distribution are estimated according to the maximum likelihood of the pixels in a non-local region. The difference between two pixels is calculated by the L2 norm distance corresponding to the Poisson distribution. Moreover, the similarity weights are defined by the sum of L2 norms between two pixels in their neighborhood points. The image Poisson denoising is carried out by using the principle of non-local mean. The experimental results show that the proposed algorithm can preserve the image details well, and the denoised image has a high peak signal-to-noise rate over 22 dB. The proposed method can be effectively used for image denoising.