复Daubechies小波域HMT模型Bayesian图像去噪  

Bayesian image denoising using HMT models in complex Daubechies wavelet domain

作　　者：褚标[1] 李昕[2] 朱功勤[1] 汪金菊[1] 

机构地区：[1]合肥工业大学计算机与信息学院,合肥230009 [2]安徽电视台制作部,合肥230022

出　　处：《计算机应用研究》2008年第4期1103-1104,1110,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(10171026)

摘　　要：提出了基于复Daubechies小波域隐马尔可夫树(SDW-HMT)模型Bayesian图像去噪算法,由于SDW小波是紧支撑、对称、正交小波,且具有近似线性相位,将其与HMT模型结合,能够更加准确地刻画小波系数的统计特征,在Bayesian图像去噪中获得很好的效果,仿真实例显示了所提算法的有效性。This paper presented a Bayesian image denoising algorithm based on hidden Markov trees models in complex Daubechies wavelet domain. The complex Daubechies wavelet had some good properties such as compact support, orthogonality, symmetry and approximate linear phase, The HMT models in complex Daubechies wavelet domain capture the key features of the joint statistics of the wavelet coefficients of image. The experimental results show the presented image denoising algorithm is effective.

关 键 词：复Daubechies小波 隐马尔可夫树模型 线性相位 贝叶斯去噪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]