Verification by Actual Measurement of Theoretical Relationship between Standard Deviation and Wiener Spectrum in Noise Analysis for Digital Radiographic Systems: Verification in One- and Two-Dimensional Wiener Spectrum
Objective: This study evaluates image noise using two theoretical equations:: standard deviation (σ) as a spatial domain analysis method and Wiener spectrum (WS) as a spatial frequency domain analysis method. The distinction between the two theoretical equations is that σ2 corresponds to the area of the one-dimensional (lD) WS and the volume of the two-dimensional (2D) WS. The aim was to investigate these equations using actual X-ray images. Methods: The X-ray imaging system employed was an AeroDR2 1417HQ (Konica Minolta Japan, Tokyo, Japan), featuring an indirect conversion-type flat panel detector. The 1D WS was obtained using the virtual slit method, while the 2D WS was obtained through the 2D Fourier transform method. Results: The relative error between σ2 measured by the virtual slit size and the area of the 1D WS was less than 0.40%. The relative error between σ2 measured by pixel size and the volume of the 2D WS was close to 0.00%. Conclusion: The accuracy of both theoretical equations was confirmed with a degree of agreement exceeding 99.5%. These minimal relative errors validated the accuracy of the theoretical equations. Given the requirement for analytical values in 2D images, we propose adopting the equation where σ2 equals the volume of the 2D WS rather than the equation where σ2 equals the area of the 1D WS.
