Signal statistics in x-ray computed tomography

摘要: Probability distribution functions (pdfs) of x-ray computed tomography (CT) signals form the basis for statistical reconstruction algorithms and noise-simulation experiments. The conventional model pdf's assumes a quanta-counting process obeying discrete Poisson distribution. In reality, CT scanners employ energy-integrating sensors detecting polyenergetic X-ray beam, with data quantized digital reconstructions. A was developed signal consisting an energy spectrum quanta (individually statistics), contributing incremental signal, proportional to their energy, analog-to-digital converter. Using moment generating function approach, pdf is shown be Compound that functionally dependent on spectrum, flux level, quantization step size. Pdfs were numerically by Fourier transform characteristic compared experimental pdfs collected from phantom scans. For exposures encountered in normal clinical usage, similar Gaussian approximation, rescaling spectra. low intensities, deviation photon-counting (Poisson) significant may have implications algorithms.