NPSIF in uniform image conditions

Figures 4 and 5 show the results under ultra-low-dose conditions. The noise-reduction processing significantly improved image granularity, as demonstrated by an overall decrease in NPS values. Furthermore, evaluations based on NPSIF revealed that the noise reduction effect on uniform images was consistent across all spatial frequencies, with particularly strong improvements observed under the INR10 condition.

Fig 4: Differences in appearance of uniform noise images in various processing: ultra-low-dose condition (0.5 mAs). Window Level is mean pixel value, and Window Width is 100.

Fig 5: Comparison of NPS characteristics and calculation results of NPSIF(u) in various processing: ultra-low-dose condition (0.5 mAs).

Figures 6 and 7 show the results under moderate-low-dose conditions. The trends of NPS and NPSIF were similar to those for the previous condition, confirming the consistent effectiveness of noise reduction processing.

Fig 6: Differences in appearance of uniform noise images in various processing: moderate-low-dose condition (1.0 mAs). Window Level is mean pixel value, and Window Width is 100.

Fig 7: Comparison of NPS characteristics and calculation results of NPSIF(u) in various processing : moderate-low-dose condition (1.0 mAs).

Figures 8 and 9 show the results under normal-dose conditions. Under this condition, the distributions of NPS exhibited a different tendency from the two lower-dose conditions. The NPSIF assessment demonstrated that the noise reduction effect of INR processing was less influential compared to conventional processing. Especially, under the INR10 condition, NPSIF tended to decrease with increasing spatial frequencies.

Fig 8: Differences in appearance of uniform noise images in various processing: normal-dose condition (5.0 mAs). Window Level is mean pixel value, and Window Width is 100.

Fig 9: Comparison of NPS characteristics and calculation results of NPSIF(u) in various processing : normal-dose condition (5.0 mAs).

Comprehensive comparison of NPSIF across all dose conditions revealed that while the improvement effect of INR processing varied depending on the dose condition, Con-NR processing exhibited a consistent effect regardless of the dose condition.

Fig 10: Comprehensive comparison of NPSIF(u) across all dose conditions in the analysis of uniform images.

NPSIF in phantom image conditions

Figures 11 and 12 show the results under low-dose conditions. It was confirmed that noise reduction processing decreased the NPS even in regions containing complex anatomical structures, such as the lung field and mediastinum. However, when evaluating the effect using NPSIF, a non-constant change along the spatial frequency axis was observed, differing from the improvement in uniform images.

Fig 11: Differences in appearance of phantom images in various processing: low-dose condition (0.5 mAs). Window Level is 2000, and Window Width is 3000 for phantom images. Window Level is 0, and Window Width is 300 for noise images.

Fig 12: Comparison of NPS characteristics and calculation results of NPSIF(u) in various processing : low-dose condition (0.5 mAs).

Figures 13 and 14 show the results under standard-dose conditions. The trend of change in NPS was similar to that observed under lower-dose conditions, but the NPS values themselves did not exhibit significant changes with INR processing. The NPSIF of INR processing showed a more pronounced dependency on spatial frequency, with clear variations observed across frequencies.

Fig 13: Differences in appearance of phantom images in various processing: standard-dose condition (1.0 mAs). Window Level is mean pixel value, and Window Width is 3000 for phantom images. Window Level is 0, and Window Width is 300 for noise images.

Fig 14: Comparison of NPS characteristics and calculation results of NPSIF(u) in various processing : standard-dose condition (1.0 mAs).

Figures 15 and 16 show the results under high-dose conditions. In this situation, there were minimal changes in NPS values due to noise reduction processing. The evaluation of NPSIF revealed that while the improvement effect of conventional processing was smaller, the frequency-dependent changes followed a trend similar to other dose conditions. On the other hand, with INR processing, the NPSIF remained nearly constant across spatial frequencies, indicating minimal improvement in NPS.

Fig 15: Differences in appearance of phantom images in various processing: high-dose condition (5.0 mAs). Window Level is mean pixel value, and Window Width is 3000 for phantom images. Window Level is 0, and Window Width is 300 for noise images.

Fig 16: Comparison of NPS characteristics and calculation results of NPSIF(u) in various processing : high-dose condition (5.0 mAs).

Comparing all dose conditions, there was a tendency for the NPSIF to increase as the spatial frequency increased with conventional processing. Conversely, for INR processing, an improvement effect was observed up to a certain spatial frequency, beyond which a constant value was maintained, highlighting a distinct characteristic.

Fig 17: Comprehensive comparison of NPSIF(u) across all dose conditions in the analysis of phantom images..