An adaptive image denoising method for mixture Gaussian noise
Yiwen Qiu, Zongliang Gan, Yaqiong Fan, Xiuchang Zhu · 2011
In real applications, different regions in the image will often be corrupted by different noise. This paper discusses this problem and proposes an adaptive image denoising method for mixture Gaussian noise in an image. There are two phases in our proposed method: noise estimation and noise removal. To estimate different noise standard deviations in different regions, a high-pass filter using two-order difference is applied first. Then we divide the filtered image into a number of blocks and compute the histogram of all the blocks' standard deviations. An averaging over the peak will estimate the noise standard deviation in one region. Finally, we combine data fusion with an existent denoising method to remove noise. Experiments with test images show that our proposed method leads to good results.