期刊文献+

基于余弦相似度的非局部均值滤波方法 被引量:9

Non-local Means Filter Method Based on Cosine Similarity Index
下载PDF
导出
摘要 为了解决非局部均值滤波(NLM)中会出现过度滤波,模糊了边缘结构信息等问题,提出了一种基于余弦相似度非局部均值滤波方法。该方法用余弦相似度改进非局部均值滤波中子块相似度的度量,能利用结构信息,对图像边缘结构信息进行更好的保持,同时可以减少图像明暗程度对去噪效果的影响。通过多个典型图像和不同的滤波参数h的实验表明,该算法与经典非局部均值滤波算法、基于积分图像的非局部均值滤波算法、Adaptive Wavelet Threshold算法、2VAR-BMWP-MAP算法、减小斑点扩散算法相比,实验结果表明:该算法能在有效去除噪声的同时更好保持边缘结构信息。另外,针对少有图像评价指标能在反映图像去噪程度的同时反映去噪算法的细节保持程度,在方法噪声的基础上提出了一种新的图像去噪评价指标,定义为方法噪声差(CB)。结论表明:方法噪声差的确能反应去噪程度的同时反应图像细节的保持程度,且比误差的均方差(MSE)更符合人的主观视觉感受。 For the non-local means filter(NLM),excessive filtered occurs,and the edge structure information is blurred.A non-local means filter method based on cosine similarity is proposed.This method improves the similarity measure of subblocks in non-local mean filtering by using cosine similarity.It can use structure information to better preserve the image edge structure information,and can reduce the influence of image brightness and darkness on denoising effect.Experiments with multiple typical images and different filtering parameters h show that the algorithm is combined with classical non-local mean filtering algorithm,non-local mean filtering algorithm based on integral image,Adaptive Wavelet Threshold algorithm,2VAR-BMWP-MAP algorithm,,the algorithm can better preserve the edge structure information while effectively removing noise.In addition,for the rare image evaluation index,the degree of detail preservation of the denoising algorithm can be reflected while reflecting the degree of image denoising.Based on the method noise,a new image denoising evaluation index is proposed,which is defined as the method noise difference.CB).Experiments show that the method noise difference can reflect the degree of denoising and the degree of detail retention of the image,and it is more in line with the subjective visual perception than the mean square error(MSE)of the error.
作者 刘传义 王世峰 王开鑫 陈森 孙琪 LIU Chuan-yi;WANG Shi-feng;WANG Kai-xin;CHEN Sen;SUN Qi(School of Optoelectronic Engineering,Changchun University of Science and Technology,Changchun 130022)
出处 《长春理工大学学报(自然科学版)》 2021年第2期18-26,共9页 Journal of Changchun University of Science and Technology(Natural Science Edition)
基金 吉林省自然科学基金项目(20150101047JC)。
关键词 图像去噪 图像评价 细节保持 非局部均值滤波 余弦相似度 image denoise image evaluation detail preservation non-local means filter cosine similarity
  • 相关文献

参考文献5

二级参考文献36

  • 1付树军,阮秋琦,李玉,王文洽.基于各向异性扩散方程的超声图像去噪与边缘增强[J].电子学报,2005,33(7):1191-1195. 被引量:22
  • 2刘芳,刘文学,焦李成.基于复小波邻域隐马尔科夫模型的图像去噪[J].电子学报,2005,33(7):1284-1287. 被引量:13
  • 3Eng H L, Ma K K. Noise adaptive soft-switching median filter for image denoising [ A ] . In: Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing [ C ], Istanbul, Turkey, 2000: 2175-2178.
  • 4Khryashchev V V, Apalkov I V, Priorov A L,et al. Image denoising using adaptive switching median filter [ A ]. In: Proceedings of IEEE International Conference on Image Processing [ C ] , Genova, Italy, 2005: 117-120.
  • 5Yang P, Basir O A. Adaptive weighted median filter using local entropy for ultrasonic image de-noising [ A ]. In: Proceedings of the 3rd International Symposium on Image and Signal Processing and Analysis [ C] ,Rome, Italy, 2003: 799-803.
  • 6Chen Y, Han C. Adaptive wavelet threshold for image denoising [J]. Electronics Letters, 2005, 41(10) : 586-587.
  • 7Donoho D L. De-noising by soft-thresholding [ J]. IEEE Transactions on Information Theory, 1995, 41 (3) : 613-627.
  • 8Chang S G, Yu B, Vetterli M. Adaptive wavelet thresholding for image denoising and compression [ J ]. IEEE Transactions on Image Processing, 2000, 9(9): 1532-1546.
  • 9Portilla J. Full blind denoising through noise covariance estimation using gaussian scale mixtures in the wavelet domain [ A ]. In: International Conference on Image Processing [ C ], Singapore, 2004 : 1217-1220.
  • 10Samsonov A A, Johnson C R. Noise-adaptive anisotropic diffusion filtering of MR1 images reconstructed by SENSE (sensitivity encoding) method [ A ]. In: Proceedings of IEEE International Symposium on Biomedical Imaging: Macro to Nano [ C ], Washington DC, USA, 2002:701-704.

共引文献78

同被引文献95

引证文献9

二级引证文献25

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部