计算机工程与应用 ›› 2021, Vol. 57 ›› Issue (20): 203-209.DOI: 10.3778/j.issn.1002-8331.2006-0417

• 图形图像处理 • 上一篇    下一篇

自适应广义全变差的图像泊松去噪算法

王洁,金正猛,冯灿   

  1. 1.南京邮电大学 理学院,南京 210023
    2.北方信息控制研究院集团有限公司,南京 211153
  • 出版日期:2021-10-15 发布日期:2021-10-21

Adaptive Generalized Total Variation Algorithm for Poisson Noise Removal

WANG Jie, JIN Zhengmeng, FENG Can   

  1. 1.School of Science, Nanjing University of Posts and Telecommunications, Nanjing 210023, China
    2.Northern Information Control Research Institute Group Co., Ltd., Nanjing 211153, China
  • Online:2021-10-15 Published:2021-10-21

摘要:

针对医学、天文图像中的泊松噪声,基于广义全变差的图像泊松去噪模型,结合交替迭代极小化方法,提出一种自适应广义全变差的图像去噪算法。该算法利用广义交叉验证技术,使得模型中的正则化参数在算法迭代过程中可以自动更新。数值实验结果验证了该算法的有效性与可行性。

关键词: 图像去噪, 泊松噪声, 广义交叉验证, 交替极小化

Abstract:

Aiming at the Poisson noise in medical and astronomical images, this paper studies the image Poisson denoising model based on generalized total variation, and proposes an adaptive generalized total variation image denoising algorithm combined with alternating iterative minimization method. The algorithm uses the generalized cross validation technology, so that the regularization parameters in the model can be updated automatically in the iterative process of the algorithm. Finally, the numerical experiment results verify the effectiveness and feasibility of the algorithm.

Key words: image denoising, Poisson noise, generalized cross validation, alternating minimization