Investigation into the optimal linear time-invariant lag correction for radar artifact removal

Area of Science:

• Medical Physics
• Radiological Imaging
• Detector Technology

Background:

• Detector lag, a residual signal in amorphous silicon (a-Si) flat-panel (FP) detectors, causes shading artifacts in cone-beam computed tomography (CBCT).
• Existing correction models often use a linear, time-invariant (LTI) approach with impulse response functions (IRFs), but IRF determination methods vary.

Purpose of the Study:

• To investigate detector lag in the Varian 4030CB FP detector.
• To identify the optimal IRF measurement technique for minimizing CBCT shading artifacts.

Main Methods:

• Examined lag linearity using rising and falling step-response functions (RSRF/FSRF) at various exposure levels.
• Implemented a multiexponential LTI model for lag correction.
• Evaluated different IRF determination techniques (RSRF vs. FSRF, exposure intensity, exposure duration, spatial position).
• Applied IRFs to projection data and CBCT scans of phantoms.

Main Results:

• Observed significant nonlinearity (>50%) in FSRF data, leading to over/undercorrection when using mismatched IRF calibration.
• CBCT reconstructions showed residual errors of 3-21 HU depending on the IRF technique.
• FSRF-based techniques at low exposures (1.6-3.4% saturation) yielded the lowest average error for CBCT.

Conclusions:

• The selection of RSRF/FSRF and exposure intensity for IRF calibration significantly impacts residual lag.
• IRFs derived from FSRF data at low exposure intensities are most effective for correcting CBCT shading artifacts.
• The choice of IRF for lag correction can be optimized based on object size.