Abstract: In order to improve the status of white Gaussian noise of computed tomography images (CT images) in the imaging process, it directed an improved wavelet-shrinkage threshold denoising methods with good time-frequency localization and multi-resolution analysis capability.It's able to overcome the man-made shocks caused by singular points in the neighborhood while denoising signal reconstruction in a the traditional wavelet threshold, provide more abundant details of medical diagnosis, better edge information and better visual images.