论文部分内容阅读
基于小波阈值的图像消噪方法是简单而又有效的,对小波系数进行空间自适应的研究可使阈值能自适应于图像的统计特性,可进一步提高消噪性能。分析了现有的自适应建模方法在消噪性能和计算消耗上的不足。在假定小波系数为具有未知分布参数的广义高斯分布随机变量的基础上提出了一种基于方差的上下文局部建模方法,用于估计每个系数所对应的参数。该方法能很好地反映小波系数的局部统计特性。实验证明其消噪效果要好于其它空间自适应建模方法。
The image denoising method based on wavelet threshold is simple and effective. Studying the spatial adaptive of wavelet coefficients can make the threshold adaptive to the statistical characteristics of the image and further improve the performance of denoising. The disadvantages of the existing adaptive modeling methods in noise cancellation performance and computational cost are analyzed. Based on the assumption that the wavelet coefficients are generalized Gaussian distributed random variables with unknown distribution parameters, a local variance based context modeling method is proposed to estimate the parameters corresponding to each coefficient. This method can well reflect the local statistical properties of wavelet coefficients. Experiments show that the noise reduction effect is better than other spatial adaptive modeling methods.