计算机工程 ›› 2010, Vol. 36 ›› Issue (14): 177-178.doi: 10.3969/j.issn.1000-3428.2010.14.064

• 图形图像处理 • 上一篇    下一篇

基于复方向滤波器组和HMT的图像去噪

尚赵伟1,张 峰1,马尚君2,朱贝贝1,国 庆1   

  1. (1. 重庆大学计算机学院,重庆 400030;2. 西北工业大学机电学院,西安 710072)
  • 出版日期:2010-07-20 发布日期:2010-07-20
  • 作者简介:尚赵伟(1968-),男,副教授、博士,主研方向:计算机视觉,模式识别;张 峰、马尚君、朱贝贝、国 庆,硕士
  • 基金项目:
    国家自然科学基金资助重点项目(90820306);国家自然科学基金资助面上项目(60873092);教育部高等学校博士学科点专项科研基金资助项目(20060611009)

Image Denoising Based on Complex Directional Filter Bank and HMT

SHANG Zhao-wei1, ZHANG Feng1, MA Shang-jun2, ZHU Bei-bei1, GUO Qing1   

  1. (1. College of Computer Science, Chongqing University, Chongqing 400030;2. School of Mechatronic Engineering, Northwestern Polytechnical University, Xi’an 710072)
  • Online:2010-07-20 Published:2010-07-20

摘要: 提出一种基于金字塔对偶树方向滤波器组(PDTDFB)分解系数模的HMT模型,该模型结合PDTDFB理论、复数的模和HMT的特点,利用PDTDFB对图像分解后复系数的模建立HMT模型,由EM算法训练模型获得去噪后的模,恢复复系数、重构图像。实验结果证实,与其他几种典型的去噪算法定性比较,该模型去噪效果有不同程度的提高,更好地保留了图像的边缘信息。

关键词: 金字塔对偶树方向滤波器组, 隐马尔可夫树模型, 图像去噪

Abstract: This paper presents an HMT model based on the module of the coefficient in Pyramidal Dual-Tree Directional Filter Bank(PDTDFB). The features of the PDTDFB theory, the module of the complex coefficient and the HMT are combined in this model. The HMT model based on the module of the complex coefficient obtained by PDTDFB decomposition is established. It trains the model by the EM algorithm to denoise and obtain the module, and restores the complex coefficient and reconstructing image. Compared with the other typical denoising methods, experimental results demonstrate that the effects raise in different degrees in image denoising, especially in edge maintenance aspect.

Key words: Pyramidal Dual-Tree Directional Filter Bank(PDTDFB), Hidden Markov Tree(HMT) model, image denoising

中图分类号: