基于双树复数小波降噪的扩散张量估计

Dual-Tree Complex Wavelet Based Denoising for Diffusion Tensor Estimation

本文引入了一个从扩散加权图像序列中估计扩散张量的新过程。此过程的第一个步骤是对扩散加权图像降噪,降噪的算法基于双树复数小波变换和双变量收缩规则;第二个步骤对降噪后的扩散加权图像使用最小二乘法估计扩散张量。我们在模拟二阶张量场和真实DT-MRI数据集上进行了实验,与扩散张量估计后的平滑方法相比,先对扩散加权图像降噪能够得到更加精确的张量场估计。估计的扩散张量场的客观结果及主观评价可以说明,本文提出的估计过程可以很好地改进最终的扩散场的质量。

This paper presents introduces a new procedure to estimate the diffusion tensor from a sequence of diffusion-weighted images. The first step of this procedure consists of the denoising of the diffusion-weighted images. The noise removal algorithm relies on dual-tree complex wavelet transform and bivariate shrinkage rifles. The second step of the procedure amounts to estimating diffusion ten- sor from denoised diffusion-weighted images using the least squares method. Experiments were performed on both synthetic second-order tensor field and real DT-MRI data set. When compared to the smoothing schemes after tensor estimation, denoising the diffusion-weighted images leads to a more accurate estimated tensor field. The objective results and subjective evaluation of the estimated diffusion tensor field also prove that the proposed procedure highly improves the quality of the final diffusion tensor field.