Abstract: According to the different distribution of noises and signals under the different scales of the wavelet transform, a mixed adaptive image denosing algorithm based on context model and wavelet transform is proposed. In this paper, different thresholding methods are adopted under the different scales of the wavelet transform. The experiment results show that the proposed method is more effective than other methods both in removing image noise and in reserving the image edge. It also can improve in Peak Signal-to-Noise Ratio（PSNR） and visual quality.

Key words: image denosing, Context model, wavelet transform, self-adaptive

摘要： 根据噪声和信号的小波系数在不同分解尺度、不同方向上高频系数的分布不同，结合Context模型，提出基于Context模型的小波变换阈值自适应图像去噪算法。该算法通过对不同尺度和方向的小波分解系数应用不同的阈值方法进行去噪。实验表明，方法能较好地去除图像噪声和保留图像边缘细节信息，在提高去噪图像信噪比值和改善视觉效果方面都表现出了良好的性能。

关键词: 图像去噪, Context模型, 小波变换, 自适应