Elimination of gridlines in X-ray image

The mammography X-ray equipment uses a radiographic grid in order to remove the scatter radiation and improve image contrast. The grid causes striped lines to appear on the film and may degrade diagnostic information. In this study, we propose a new method which reduces the noise caused by the grid on mammographic film using digital image processing techniques. We considered a method of eliminating gridlines based on the mechanism of gridline appearance and investigated the validity of image filtering. First, we exposed a phantom and selected a region of gridlines for study. We smoothed gridline image based on the line profile. We detected the ripple components of the gridlines and subtracted them from the micro calcifications image. Next we performed median filtering and the V-filtering as a method of image filtering. Then, we calculated the standard deviation of the line profile and signal to noise (SN) ratios in the micro calcifications region. The standard deviations of the line profile processed by the two methods were decreased and the SN ratios of the respective images were increased. The SN ratio of the image processed by median filtering after the V-filtering was approximately equal to that of the image processed by subtracting ripple components. Thus, we could confirm the validity of image filtering. It was found out that our proposed image filtering method was useful for eliminating gridlines in the X-ray image.