Fourier Analysis of Noise Characteristics in Cone-Beam Microtomography Laboratory Scanners.

Abstract

We investigate the signal and noise performance of an x-ray microtomography system that incorporates a complementary metal-oxide-semiconductor flat-panel detector as a projection image receptor. Signal and noise performance is analyzed in the Fourier domain using modulation-transfer function (MTF), noise-power spectrum (NPS), and noise-equivalent number of quanta (NEQ) with respect to magnification and different convolution kernels for image reconstruction. Higher magnification provides lower NPS, and thus, higher NEQ performance in the transaxial planes from microtomography. A window function capable of smoothing the ramp filter edge to below one-half of the Nyquist limit results in better performance in terms of NPS and NEQ. The characteristics of convolution kernels do not affect signal and noise performance in longitudinal planes; hence, MTF performance mainly dominates the NEQ performance. The signal and noise performances investigated in this study are demonstrated with images obtained from the contrast phantom and postmortem mouse. The results of our study could be helpful in developing x-ray microtomography systems based on flat-panel detectors.