Estimation Of Signal Dependent Noise Parameters From A Single Image

Xinhao Liu, Masayuki Tanaka and Masatoshi Okutomi

Abstract

     The additive white Gaussian noise (AWGN) is usually assumed in many image processing algorithms. However, these algorithms cannot effectively deal with the noise from actual cameras which is better modeled as signal dependent noise(SDN). In this paper, we focus on the SDN model and propose an algorithm to accurately estimate its parameters without any assumption of the noise types. The noise parameters are estimated by using the selected weak textured patches from a single noisy image. Experiments on synthetic noisy images are conducted to test the algorithm, which show that our noise parameter estimation outperforms the existing algorithms. And based on our estimation, the performance of image processing applications like Wiener filter can be effectively improved.

Signal Dependent Noise Model

where g is the noisy pixel value, f is the noise-free pixel value, γ is the exponential parameter, and u and w are zero-mean random variables with variances σ_u and σ_w ,

Experiment Results

1) Noise Parameter Estimation On Berkeley Segmentation Dataset

2) Denoising Performance Based on Estimated Noise Parameter:

Download

Matlab [code]

Reference

1. Estimation of Signal Dependent Noise Parameters from a Single Image [pdf]
   Xinhao Liu, Masayuki Tanaka and Masatoshi Okutomi
   Proceedings of IEEE International Conference on Image Processing (ICIP2013), September, 2013