基于非降采样 Contourlet变换的非线性图像增强新算法

张林; 朱兆达

doi:10.3724/SP.J.1146.2008.01025

基于非降采样 Contourlet变换的非线性图像增强新算法

doi: 10.3724/SP.J.1146.2008.01025

张林,
朱兆达

计量
- 文章访问数: 3820
- HTML全文浏览量: 108
- PDF下载量: 982
- 被引次数: 0
出版历程
- 收稿日期: 2008-08-26
- 修回日期: 2009-03-05
- 刊出日期: 2009-08-19

A Novel Nonlinear Method for Image Enhancement Based on Nonsubsampled Contourlet Transform

摘要

摘要: 为了克服传统去图像噪声算法的限制，该文提出一种基于非降采样(Nonsubsampled)Contourlet变换的增强新算法(NNIEM-NSCT)。此新算法通过充分利用方向子带相关性的自适应贝叶斯阈值，既保护了图像边缘细节，又可更好地抑制图像噪声。其次，文中构造的非线性增强匹配函数，通过改变变换域的系数能有效对图像强弱边缘进行不同程度的增强。实验结果证明，该文新算法在图像细节处理上，优于基于NSCT的方法，细节方差( DV) 大约为NSCT的2倍，背景方差(BV)基本保持不变，并且具有更好的视觉效果。
- 图像增强;非线性增强函数;非降采样Contourlet变换(NSCT);自适应贝叶斯阈值
Abstract: To overcome the limitation of traditional image denoising algorithms, a Novel Nonlinear Image Enhancement Method based on NonSubsampled Contourlet Tansform (NNIEM-NSCT) is proposed in this paper. The method not only preserves the details of image edge but also suppresses the noise by means of an adaptive Bayes shrinkage threshold which takes advantages of the relationship across all subbands in the same level. Secondly, a nonlinear enhancement mapping function is constructed by modifying the NSCT coefficients to change the image enhancement of weak and strong edges efficiently. Experimental results show that the proposed method outperforms the NSCT-based method. The Detail Variance (DV) of proposed method improves about two times more than that of NSCT-based method, when the Background Variance (BV) is almost equal, and a better visual effect is gotten.

HTML全文

参考文献(1)

Dippel S, Stahl M, Wiemker R, and Blaffert T. Multiscalecontrast enhancement for radiographies: Laplacian pyramidversus fast wavelet transform. IEEE Transactions on MedicalImaging, 2002, 21(4): 343-353.[2]Po D D and Do M N. Directional multiscale modeling ofimages using the contourlet transform[J].IEEE Transactionson Image Processing.2006, 15(6):1610-1620[3]何宏, 唐志航等. 基于小波多尺度积的图像增强新算法. 计算机应用与软件, 2007, 24(3): 163-165.He Hong and Tang Zhi-hang, et al.. Novel algorithm for imageenhancement based on wavelet multiscale product. ComputerApplication and Software, 2007, 24(3): 163-165.[4]Do M N and Vetterli M. The contourlet transform: Anefficient directional multiresolution image representation[J].IEEE Transactions on Image Processing.2005, 14(12):2091-2106[5]梁栋, 沈敏等. 一种基于Contourlet递归Cycle Spinning的图像去噪方法. 电子学报, 2005, 33(11): 2044-2046.Liang Dong and Shen Min, et al.. A method for imagedenoising based on the contourlet transform using recursivecycle spinning. Acta Electronica Sinica, 2005, 33(11):2044-2046.[6]Zhou J, Cunha A L, and Do M N. Nonsubsampled contourlettransform: construction and application in enhancement.Proc. of IEEE Intl Conf on Image Processing, Genoa, Italy,Sep. 2005: 469-472.[7]Cunha A L, Zhou J, and Do M N. The nonsubsampledcontourlet transform: theory, design and applications[J].IEEETransactions on Image Processing.2006, 15(10):3089-3101[8]Feng Peng, Pan Ying-jun, Wei Biao, Jin Wei, and Mi De-ling.Enhancing retinal image by the Contourlet transform[J].Pattern Recognition Letters.2007, 28(4):516-522[9]Khan M A U, Khan M K, and Khan M A. Coronaryangiogram image enhancement using decimation-freedirectional filter banks. Proc. Int. Conf. Acoutics, Speech andSignal Proc.(ICASSP), Montreal, QC, Canada, 2004:441-444.[10]Chang S G, Yu B, and Vetterli M. Spatially adaptive waveletthresholding with context modeling for image denoising[J].IEEE Transactions on Image Processing.2000, 9(9):1522-1531[11]Da C and Minh N D. Nonsubsampled contourilet transform:filter design and applications in denoising[C]. IEEEInternational Conference on Image Processing, Genova, Italy,2005: 749-752.[12]Nezhadarya E and Shamsollahi M B. Image contrastenhancement by Contourlet transform. 48th InternationalSymposium ELMAR-2006 focused on Multimedia SignalProcessing and Communications, June 2006: 81-84.

施引文献

资源附件(0)

访问统计