Noise power spectrum of computed mammography and full field digital mammography

Abstract

Noise in medical imaging is one of the parameters that need to be measured in image analysis. Noise of an image can be measured by noise power spectrum (NPS). A comparison of two types of mammography imaging machine namely Hologic Lorad Selenia Full-field Digital Mammography (FFDM) and Hologic Lorad M-IV Series Computed Mammography (CM) has been made to identify the system of mammography that has higher noise level. Both types of machine use automatic exposure control (AEC) for image acquisition. A potential of 30 kVp and 110 mAs-120 mAs has been applied. The images that are obtained from both of machines are analysed using MATLAB programme MATLAB Version 7.8.0 (R2009a). The best technique of NPS computations, Dobbin’s method has been applied. NPS is shown using 2-Dimensional mesh plot and 1-Dimensional graph taken from the extraction of 2-Dimensional graph where the noise power spectra versus values spatial frequency are shown. Noise level of the images from both machines are studied and compared. The sensitivity of NPS is at lower spatial frequency.The comparison shows that at lower spatial frequency, CM image has higher noise level compared to FFDM image, and at higher frequency, FFDM image has higher noise level compared to CM image. The noise level is in the range of 10-3-10-6 mm2. In conclusion, CM has higher noise compared to FFDM.