Multiscale Wiener Filtering Method For Low-Dose CT Images

Degradation of low-dose CT image such as noising and detail blurring can worsen visual quality and the results of analysis, so it is necessary to enhance low-dose CT image to improve its visual quality. In this paper, a multi-scale Wiener filtering method is proposed based on analysis of low-dose CT image noise. First, the algorithm decomposes image by two-stage wavelet transformation, and then design Wiener filters for the every sub-images. The value of the peak signal to noise ratio (PSNR) of low-dose CT images was used as the measure of image visual quality. Experiment results demonstrate that the algorithm improves the image visual quality. It can get the conclusion that the method denoised low-dose CT image with good visual effect, and it can be used to denoise low-dose of CT images in medical field.