Optimizing PET Detector Performance and Calibration through Flood Histogram Analysis

In Positron Emission Tomography (PET) detectors, effective analysis of flood histograms is essential for evaluating detector performance. This study outlines a detailed procedure for generating and analyzing these histograms, using techniques including Weighted Spot Deviation (WSD), Voronoi Diagrams (VD), and Delaunay Triangulation (DT). We applied these techniques to a PET detector system utilizing an 8x8 LSO scintillator crystal array and a data acquisition circuit integrating a Sigma-Delta ADC and FPGA-based DAQ. The WSD metric was applied in a case study where flood histograms were generated from a 4x4 SensL SiPM array processing 16 individual signals, and a PSPMT processing 64 signals multiplexed into four output signals. Detailed analyses, including position profiling, centroid connectivity, anomaly detection, and spatial analysis through segmentation, skeletonization, and triangulation, were conducted. These methods enable a comprehensive assessment of detector performance through spatial analysis, identify nonuniformities, and improve detector calibration, thus enhancing quality control and characterization of PET detectors.