Verification of the influence of the sampling aperture on the digital noise power spectrum

Congress:

ECR 2020

Poster Number:

C-04596

Type:

Scientific Exhibit

Keywords:

Image Quality, Radiation physics, Conventional radiography, Digital radiography, Experimental investigations, Physics, Image verification, Not applicable, Experimental, Performed at one institution

Authors:

S. Maruyama¹, M. Shimosegawa²; ¹Takasaki-shi, Gunma/JP, ²Maebashi-shi/JP

DOI:

10.26044/ecr2020/C-04596

DOI-Link:

https://dx.doi.org/10.26044/ecr2020/C-04596

Fig. 1: Definitions of measures used in this study.

Fig. 2: Geometry of the simulation used for calculating the NPS.

Fig. 3: Geometry of the experimental measurement used for calculating the NPS.

Fig. 4: Definitions of the correction factor used in this study.

Fig. 5: Outline of verification in experimental measurement. The NPS can be corrected...

Methods and materials

Simulation

An image formation simulation was performed, assuming a direct-type FPD with ideal X-ray absorption characteristics, such that it has no unsharpness component other than the sampling aperture, and no noise factors other than quantum mottle [10,11]. In this study, we applied a Monte Carlo simulation: Electron Gamma Shower Ver.5 [12].

Figure 1 shows the definitions of measures used in this study. To verify the aperture size dependency of the NPS_digital(u), the sampling aperture was 0.02, 0.05, 0.1, and 0.2 mm, and X-ray photons were uniformly incident with 10⁵ mm^-2 incident on the detector (Figure 2). the NPS_digital(u) was measured from the obtained simulation image, using the two-dimensional fast Fourier transform (2D-FFT) method recommended in IEC612220-1 [7,13].

If the only unsharpness component of the detector is the aperture, the presampled MTF(u) corresponds with the sinc function. Therefore, to confirm that the image formation simulation was properly performed, a 1 mm thick tungsten edge tilted 2.5° with respect to the pixel alignment is placed in the simulation geometry used for NPS measurement. From this, simulation edge images were acquired. Then, the presampled MTF(u) at each aperture size was measured using the edge method [14], and compared with the sinc function which is a true value [15].

Experimental measurement

For verification, CALNEO Smart (Fujifilm) and AeroDR (Konica Minolta) indirect-type FPDs with a CsI sensor were used. For the calculation of the presampled MTF(u), the edge method was used in the same manner as in the simulation. The 2D-FFT method was also used for the calculation of NPS_digital(u) (figure 3). The detector exposure was approximately 1 mR in RQA5 specified in IEC61627 [16].

To examine the influence of each MTF component, a coefficient for correcting the unsharpness factor due to the analog and aperture components were calculated by formula shown in figure 4.

The NPS can be corrected by multiplying these correction factor, the influence of each MTF component on NPS_digital(u) can be evaluated (figure 5). The definition of the presampled MTF is the product of MTF_analog(u) and MTF_aperture(u) [3].