36. Optimization of the acquisition threshold of Photon Counting Detectors (PCDs) used in X-ray medical imaging

2018 
Purpose PCDs with energy discrimination led to new possibilities of improving the quality of X-ray medical images. PCDs are direct detection devices in which the charge delivered by a single interacting photon is directly collected and discriminated at single pixels level. The limit of high densely pixelated PCDs is the charge sharing which can lead to multiple counts from a single interaction degrading the spatial resolution of the system. Increasing the discrimination threshold, multiple counts can be partially or totally removed from images improving the spatial resolution at the cost of a reduction of the detection area of single pixels with a consequent loss of statistics. This work aims to show a simple method to find the optimal threshold which gives the best trade-off between spatial resolution and statistics. Methods The acquisitions have been performed at the SYRMEP beamline (Elettra Trieste) using the tomographic experimental set-up of INFN project SYnchrotron Radiation MAmmography (SYRMA-3D). SYRMA-3D uses ‘Pixirad-8’, a PCD with CdTe sensor and a pixel size of 60 μm. For this study, we used a monochromatic beam of 38 keV acquiring a set of images of an edge and a set of images of the beam with different detection thresholds. The two sets of images have been used respectively for the measurement of the Modulation transfer function (MTF) and the noise power spectrum (NPS). Combining both NPS and MTF the Number of Equivalent Quanta (NEQ) can be determined. The NEQ summarizes both informations about spatial resolution and noise and is an useful metrics for the global image quality assessment. Results The comparison between NEQ curves shows that lower thresholds give better NEQ response. Conclusions The increase of the discrimination threshold improves the spatial resolution but worsens the noise. The global image quality, measured using NEQ, is higher for lower thresholds.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []