Image denoising using two-dimensional stochastic resonance in wavelet domain

It has been found that stochastic resonance can filter out the image noise and enhance the image. However, since the stochastic resonance has a characteristic that the high frequency energy is shifted to the low frequency, the denoised image will lose detail and become blurred. Because the noise and detail of the image are mainly distributed in the high frequency band, image denoising and detail retention can be regarded as the processing of high frequency. This paper applies the frequency segmentation feature of wavelet transform to decompose gray image into wavelet domain, and the high frequency part is processed by two-dimensional stochastic resonance, the processed high frequency subband and the original low frequency subband are subjected to two-dimensional inverse wavelet transform to obtain the denoised image. The purpose is to remove the image noise, and preserve the image details at the same time. In this paper, the mean square displacement(MSD) is used as the index to measure the amount of noise removal, and the edge sharpness (EAV) is used as an index to measure the image clarity. The experimental results show that the MSD and EAV of the algorithm are superior to those of the traditional noise reduction techniques.