基于非局部贝叶斯的泊松图像去噪算法

A Nonlocal Bayesian Image Poisson Denoising Algorithm

针对图像泊松噪声去除问题,提出了一种基于非局部贝叶斯的图像去噪算法,主要根据噪声模型建立相似块组的统计模型.在Bayesian-MAP框架下,利用图像的自相似性,建立基于图像块的泊松图像去噪模型,然后利用分裂Bregman法对模型进行迭代求解.实验结果验证了模型的有效性,与其他泊松去噪算法相比,模型的恢复性能在客观评价指标上具有明显的改善.

Concerning the removal problem of image Poisson noise,a novel image denoising algorithm based on nonlocal Bayesian is proposed.The main idea is to establish statistical model for a set of similar patches according to Poisson noise model.In the framework of Bayesian-MAP estimation,image self-similarity is utilized to build the patch-based image Poisson denoising optimization mode.The optimization model could be efficiently solved with splitting Bregman method.Experimental results show that the proposed method is effective and it outperforms than other methods on objective criterion.