X-ray medical imaging. X-detectors characterization

This article presents the theoretical foundations of X-ray detector characterization techniques. After a brief but essential reminder of Fourier analysis, it introduces the notion of MTF (Modulation Transfer Function), which measures the resolution performance of a detector, by explaining the contributions of its various components (scintillator and pixel matrix in particular). Next, the notions of quantum and electronic noise are presented, along with the parameter that synthesizes them (the NPS or Wiener function). The various contributions to the NPS (Noise Power Spectrum) are explained, including those linked to the effects of oversampling and spectrum folding in the case of a pixelated detector.

A third section defines the notion of signal-to-noise ratio, initially reduced to an elementary pixel. It is generalized in the fourth part, when NEQ (Noise-Equivalent Quanta) and DQE (Detective Quantum Efficiency) are introduced. The latter is now a universally used factor of merit for quantifying the performance of a digital detector, and measures nothing less than the degradation of the signal-to-noise ratio induced by the detector as a function of spatial frequency. In other words, it quantifies the loss of information attributable to the detector or, to put it another way, it estimates the dose lost at detector level, which will have to be compensated for by a surplus dose at patient level to achieve a given image quality: the higher the QED, the lower the dose to which the patient will be exposed.

At the very end of this article, other important parameters are also briefly introduced, in particular reading speed.

Flat detector technologies, whether direct or indirect detection, are the main theme of this article, but all the concepts presented can be applied to any type of X-ray detector, even if some formulas have to be adapted here and there.