Figure 3.

Plots of ${\text{NPS}}_{\text{horizontal}}$ and ${\text{NPS}}_{\text{vertical}}$ at ${\text{CTDI}}_{\text{vol}}$ of 5.81 mGy with Veo 3.0 ($\text{TT}\hspace{0.17em}2.5\hspace{0.17em}\text{mm}+\text{NR}05+\text{Texture}$) and FBP. In addition to their differences in ${\text{NPS}}_{\text{horizontal}}$ peak frequencies, the noises with Veo 3.0 had lower magnitude and distributed in a narrower range, in comparison to the FBP. In general, the noises increased when ROIs moved towards the center. There exists the exception of noise on the most outer ring with a 180‐pixel radius being higher at some frequencies than the ones on the inner rings. These discrepancies may be caused by 1) match of the phantom shape (elliptical section instead of circular) and size with the bowtie filter, 2) types of materials in the samples, and 3) anisotropic sampling: this ring samples only the lateral regions of a circle, instead of a complete circle. However, further investigation may be needed to understand the discrepancies.