Abstract: Aimed at the deficiency of universal thresholding denoising, this paper nitoroduees a new subband-adaptive denoising algorithm based on weighted threshold. After analyzing statistics of wavelet coefficient for wavelet denoising of natural image, an exponentially decaying inter-scale model for image wavelet coefficients is suggested for the adaptation of the denoising filter across scales. Then a subband-adaptive thresholding algorithm is proposed based on statistics comparison of noisy image and original one, which avoids unbalanced denoising cross scales. Experiments show that higher peak-signal-to-noise ratio and better subjective visual effect can be obtained as compared to other thresholding-denoising algorithms.

Key words: Image denoising, wavelet transform, coefficient model, weighted threshold

摘要： 针对小波全局阈值去噪的缺点，介绍了一种子带自适应的阈值加权算法。通过对图像小波分解系数统计特性的分析，提出了一种近指数模型作为分解层之间小波系数的先验分布。在此基础上，对比噪声图像和无噪图像在各尺度下统计特性，给出了一种子带自适应的加权阈值计算方法，避免了各层子带去噪的不平衡。实验表明，与全局阈值和其它子带自适应阈值去噪方法相比，基于加权阈值的图像去噪方法能获得更高的信噪比和更好的视觉效果。

关键词: 图像去噪, 小波变换, 系数模型, 加权阈值

CLC Number: 

• TP18