摘要
Contourlet变换是继小波变换之后的又一新变换。由于contourlet变换的多尺度和多方向特性,能有效地捕获到自然图像中的轮廓,并对其进行稀疏表示。详细分析了图像contourlet系数的统计特性,并利用非高斯双变量分布对系数层间相关性进行建模。最后,将此分布应用于图像去噪,就PSNR、NMSE和视觉质量这三方面的评价指标与contourlet HMT和小波阈值法进行了比较。实验结果表明:算法能获得较好的结果,尤其是对于含有丰富纹理的图像。
The contourlet transform is a new extension of the wavelet transform in two dimension.Because of its multiscale and directional properties, the contourlet transform can effectively capture the smooth contours that are the dominant features in natural images with only a small number of coefficients. The statistics of the contourlet coefficients of natural images in detail and model contourlet coefficients using non-Gaussian bivariate distribution that captures their interscale dependencies are studied. In the end,this model was tested for the image denoising. It was also compared with contourlet HMT and wavelet thresholding using PSNR, NMSE and visual quality. The results show that great performance improvements over other methods, especially for the images that have abundant texture.
出处
《红外与激光工程》
EI
CSCD
北大核心
2006年第2期234-237,共4页
Infrared and Laser Engineering
基金
国家自然科学基金资助项目(60462003)
