| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 1829961 | Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment | 2008 | 13 Pages |
On the basis of the determination of statistical correlations between neighboring detector pixels, a novel method of estimating the simultaneous detection probability of pixels and the spatial resolution of pixelized detectors is proposed. The correlations are determined using noise variance measurement for isolated pixels and for the difference between neighboring pixels. The method is validated using images from two image-acquisition devices, a General Electric Senographe 2000D and a SD mammographic unit. The pixelized detector is irradiated with X-rays over its entire surface. It is shown that the simultaneous pixel detection probabilities can be estimated with an accuracy of 0.001–0.003, with an estimated systematic error of less than 0.005. The two-dimensional pre-sampled point-spread function (PSF0) is determined using a single Gaussian approximation and a sum of two Gaussian approximations. The results obtained for the pre-sampled PSF0 show that the single Gaussian approximation is not appropriate, and the sum of two Gaussian approximations providing the best fit predicts the existence of a large (∼50%) narrow component. Support for this observation can be found in the recent simulation study of columnar indirect digital detectors by Badano et al. The sampled two-dimensional PSF is determined using Monte Carlo simulation for the L-shaped, uniformly distributed acceptance function for different fill-factor values. The calculation of the pre-sampled modulation transfer function based on the estimated PSF0 shows that the observed data can be reproduced only by the single Gaussian approximation, and that when the sum of two Gaussians is used, significantly larger values are apparent in the higher-frequency region for images from both detection devices. The proposed method does not require a precisely, constructed tool. It is insensitive to beam collimation and to system physical size and may be indispensable in cases where thin absorption slits or edges are difficult to use. It could therefore be very useful for regular detector verification.
