Improving the accuracy of MTF measurement at low frequencies based on oversampled edge spread function deconvolution

Area of Science:

• Medical Imaging
• Radiographic System Analysis
• Image Quality Assessment

Background:

• Traditional slanted edge methods for measuring modulation transfer function (MTF) in radiographic systems suffer from noise amplification during numerical differentiation.
• This noise amplification limits the accuracy of MTF measurements, particularly at low frequencies, impacting diagnostic capabilities.
• Accurate low-frequency MTF is crucial for evaluating the performance of digital x-ray imaging systems.

Purpose of the Study:

• To develop and validate an improved technique for accurate low-frequency modulation transfer function (MTF) measurement in digital x-ray imaging systems.
• To address the limitations of conventional slanted edge methods in handling noise amplification at low frequencies.

Main Methods:

• Developed an edge spread function (ESF) deconvolution technique based on a degradation model of slanted edge images.
• Constructed symmetric oversampled ESFs by subtracting a shifted version of the original ESF.
• Validated the proposed technique against the conventional slanted edge method using computer simulations and experimental data from two digital radiography systems.

Main Results:

• Computer simulations demonstrated significantly lower average errors for the ESF deconvolution technique (0.11% ± 0.09% and 0.23% ± 0.14%) compared to the conventional method (0.64% ± 0.57% and 1.04% ± 0.82%) at low frequencies.
• Experimental results on digital radiography systems showed superior uncertainty performance for the proposed technique over the conventional method.
• The ESF deconvolution technique effectively mitigates noise amplification issues inherent in traditional MTF measurement methods.

Conclusions:

• The proposed ESF deconvolution technique significantly improves the accuracy of low-frequency MTF measurements in digital x-ray imaging.
• This method offers better uncertainty performance and reduced errors compared to conventional slanted edge techniques.
• The findings demonstrate a viable approach to enhance the reliability of MTF assessments for digital radiography systems.