基于盲分离的图像去噪算法研究

Image denoising based on blind source separation.

图像去噪是图像处理中一个重要而又富有挑战性的课题。已有的图像去噪算法对噪声模型作出假设,在达到假设条件时取得较好的去噪效果。这类算法不能完全去除噪声,而且会削弱图像信号的细节。把图像和噪声看作是互相独立的两个信号源,把去噪过程作为信号分离过程来处理。在初步从污染图像中估计出一个虚拟观测图像后,用基于独立分量分析的盲分离来达到去噪目的。实验结果表明,该算法相比传统的基于滤波的去噪方法,在噪声强度很大的情况下,依然能得到较好的去噪效果。

Removing noise from the original signal is still an important and challenging problem for researchers.In spite of the sophistication of the recently proposed methods,most algorithms have not yet attained a desirable level of applicability.All show an outstanding performance when the image model corresponds to the algorithm assumptions,but fails in general and creates artifacts or removes image fine structures.In this paper the image and the noise are considered as two independent sources which mixed together.If we can separate the two sources,the aim to denoise is achieved.Blind Source Separation （BSS） techniques such as Independent Component Analysis（ICA） lend themselves well to analyse such problems.A dummy observation will be estimated from the corrupted image,and an image denoising algorithm based on BSS is presented.Demonstration indicates that the proposed method gives better result compared to conventional method.