Fourier Based Segmentation of Microcalcifications in Mammograms

This paper presents a Fourier transform approach to detect microcalcifications in digital mammograms. The basic idea consists in the design of parametric Butterworth bandpass filters in the Fourier domain used to extract sharpened border like structures that correspond to detected mammography microcalcifications. Image thresholding of the filtered image is accomplished, first by homogenizing the background (fibroglandular tissue) with a median filter, after which a gamma correction is applied to change the global contrast. Second, by postprocessing the resulting image using histogram based local and global statistics we obtain a properly binarized image that emphasizes the desired objects (microcalcifications) and segmentation is completed using a sequence of morphological binary operations. Several illustrative examples taken from a clinical database are included to demonstrate the capability of the proposed approach in comparison with other edge detection techniques such as the difference of Gaussians (DoG) and the Laplacian of a Gaussian (LoG).