关键词: 图像去噪, 小波系数, 最小均方误差, 邻域窗口, 增强因子

Abstract: Wavelet denoising is a new image denoising method.Neighborhood threshold shrinkage method is one of the wavelet domain threshold image denoising methods. Its principle is that according to the square of all the wavelet coefficients in the neighborhood window,it processes the wavelet coefficients in the center of the window. An improved image denoising method of optimization in the wavelet domain was proposed. At first, the proposed method determined an optimal threshold and the window of the neighborhood using unbiased risk estimation of mean square error criterion in the wavelet domain for each subband; then introduced a detailed enhancement factor of P, and a mapping function in order to optimize the shrinkage factor for the wavelet coefficient; at last, the estimation of the true coefficients were obtained by the shrinkage estimation of wavelet coefficients. The experimental results show that the proposed method obviously outperforms the neighborhood threshold shrinkage method in the ratio of peak signal to noise. At the same time, it effectively enhances image details, and effectively improves the visual quality of the image.

Key words: image denoising, wavelet coefficients, Minimum mean-square error, neighborhood window, enhancement factor