基于稀疏表达的图像去噪方法研究

提出一种基于混合字典的图像稀疏分解去噪方法。使用小波包函数和离散余弦函数构成混合字典,采用匹配追踪算法对图像进行稀疏分解,提取含噪图像中的稀疏成分,最后利用稀疏成分进行图像重构,达到去除图像中噪声的目的。实验中与单一字典稀疏分解去噪算法进行了对比,结果表明,所提出的混合字典稀疏去噪算法可有效提取图像中的稀疏结构,改善重构图像的主客观质量。

Research on image denoising based on sparse representation

A sparse representation image denoising method based on mixed overcomplete dictionary is proposed. Firstly, overcomplete dictionary is composed by mixing wavelet packet and discrete cosine function. And then matching pursuit algorithm is used to decompose image and extract sparse components. Finally, image is reconstructed using these sparse components. By doing so, the noise in the image is reduced. In the experiment, the proposed algorithm is compared with the common used unitary dictionary sparse representation method, and the results show that the proposed method can effectively extract sparse component of the image and improve the subjective and objective image quality.