Abstract： Objective To improve the contrast between gray matter (GM) and white matter (WM) in patients undergoing plain cerebral computed tomography (CT) with nonlinear transformation using image processing technique of matrix laboratory (MATLAB).Methods Image processing technique of nonlinear transformation was established by MATLAB.Imaging data of plain cerebral CT in 38 patients wiht suspect brain diseases (16 females and 22 males) were collected from our hospital using Siemens Dual-source CT (Model definition) to determine the mean CT value of cerebral GM and WM and their pixels,followed by post- processing of DICOM images using MATLAB.DICOM images of high and low frequencies were separated via a round filter within frequency spectrum.The contrast between GM and WM was enhanced via stretching grey scales of the images,based on nonlinear transformation for modification,with the mean of cerebral GM and WM as breakover points respectively.Results CT images of above 38 cases were normal.Of 38 normal subjects,nonlinear transformation yielded an increase of 1.5Δp in the pixel (0＜Δp＜3) at both breakover points as compared with baseline level.This was associated with enhanced contrast and separationbetween GM and WM,as well as similar quality with the primitive images.Conclusion The nonlinear transformation approach for grey scale,based on MATLAB image processing software,can improve contrast between GM and WM in cerebral plain CT in a noise-free manner and can therefore be applied in capturiug DICOM images regardless of the type of CT scanners.