The performance of four methods for feature enhanced mammographic images—optimal adaptive neighborhood imaging, automatic spatial filtering, adaptive histogram equalization imal contrast pixel to pixel—are compared. Each method is te a database of 50 mammographic images, which included similar shaped nodules. A contrast parameter was used to compare the performance of each processing method. It was found that the ma trast pixel-to-pixel method gives the highest contrast value. A mamogram including a real subtle lesion is used to illustrate the use of this last contrast enhancement method in clinical diagnosis.
Read full abstract