Fig 1.
(a): Test scanner. The filled circles represent positron annihilations and the lines represent the trajectory of the resulting annihilation photons. For some annihilations, one photon of the pair is detected while the other is lost; for some other annihilations both photons are detected. (b): schematic representation of the detection model. The photons measured by the ith detector can be modelled as those originating from a source of individual photons (it represents the situation where only one photon is detected), plus photons from a source of photon pairs (both photons are detected).
Fig 2.
SW sorter: Schematic representation.
Time is represented horizontally. Uncorrelated singles are depicted as vertical lines and pairs of correlated singles are depicted with equal symbols on their tops. The TCW is represented as a bended arrow below the data flow. Sorted prompts are tagged as P and further classified into randoms, R, and trues, T. The multiple coincidence is tagged as M.
Fig 3.
Estimated random rates and random fractions.
Error bars correspond to one standard deviation. Estimates provided by the ideal method are represented as a horizontal dotted line; SP: solid line; SR: dashed line, and DW: dotted–solid line. The average singles count rate (ACR) per detector is shown in the upper x axis. The RF is represented by a grey solid line.
Fig 4.
Estimations provided by the methods, the Fano factors and NECR for the IQ phantom.
The vertical line corresponds to the NECR value used. SP: solid line; SR: dashed line, and DW: dotted–solid line. The RF is represented by a grey solid line.
Fig 5.
First row: low set; second row: high set.
Third row: low set including the degradation effects. Ideal value: horizontal line; ideal method: dotted line; SP: solid line; SR: dashed line; DW: dotted–solid line, and no compensation applied (represented by the symbol ): dot–dot–dash.
Fig 6.
Transaxial central slice of the IQ phantom at 40 iterations.
First row: low set; second row: high set; third row: images with all the degradation effects. The images in the same row share the same colour scale, which has been normalized to the pixel with the maximum intensity.
Table 1.
Regional bias computed for the hot and warm regions.
Fig 7.
Left: central transaxial and coronal slices of the simulated MOBY phantom. Right: graph of the CC achieved by the methods.
Fig 8.
Histograms of the estimated values of .