基于二进小波变换的磁共振图像去噪新算法

针对传统去噪方法在降低噪声的同时会模糊图像边缘的缺点,提出了一种基于二进小波变换的噪声抑制新算法.利用小波系数的区域相关性,将要处理的小波系数置于由它周围的系数组成的可变窗口内,由窗口内所有小波系数的值来决定该系数的处理方式.由于离散小波是非平移不变的,因而重构过程中会出现人工噪声.为了避免这个问题,本文采用了具有平移不变性的二进小渡变换.实验结果表明,该方法较之传统去噪方法有更高的去噪精度,可以有效降低噪声,同时较好地保持图像细节和边缘信息.

Noise Reduction Algorithm for MRI Images Based on Dyadic Wavelet Transform

According to the characteristics of traditional denoising method with edge information of images while reducing the noise,an algorithm based on the dyadic wavelet transform for the noise reduction is presented.The wavelet threshold method is improved by using a localized,context-based,thresholding strategy for incorporating neighbouring coefficients on each wavelet coefficient to be thresholded.Meanwhile,the translation-invariant(TI)dyadic wavelet transform is used instead of Dyadic wavelet transform(DWT),thus avoiding the artifacts caused by the non-TI reconstruction.Experiments on the clinical images show that the algorithm has a higher de-noising accuracy compared with tradition methods.The algorithm can reduce the noise and enhance the detail and edge features of Magnetic resonance imaging(MRI)images.