Abstract
X-ray CT projection data often include components with frequencies that are markedly higher than the pixel Nyquist frequency fPN, which is determined by the pixel size. Noise components higher than fPN are folded back into a region lower than fPN through the backprojection process, thereby creating aliased noise. With clinical CT scanners, we evaluated the aliased noise using an aliasing prevention measure, band-limiting processing (BLP), which suppresses frequency components higher than fPN in the projection data. Indices we used to evaluate improvement by BLP were the noise power spectrum (NPS), modulation transfer function (MTF), signal-to-noise-ratio (SNR) spectrum, matched filter SNR (MF SNR), and two-alternative forced-choice (2-AFC) test. With BLP, the NPS was decreased not only beyond fPN, but also within fPN. The same level of MTF was maintained as that without BLP within fPN. No remarkable reduction in spatial resolution was observed. The SNR spectrum and the MF SNR of the BLP image nearly agreed with those of an ideal state without aliased noise. A notable improvement in the visuoperceptual image quality by BLP was recognized with a reconstruction field of view (FOV) of more than 45 cm. We then applied BLP to clinical data and confirmed that significant aliased noise of a large FOV image was removed without notable side effects. The results showed that at least some CTs suffering from aliased noise can be improved by proper band-limiting.
Original language | English |
---|---|
Pages (from-to) | 178-192 |
Number of pages | 15 |
Journal | Radiological Physics and Technology |
Volume | 8 |
Issue number | 2 |
DOIs | |
Publication status | Published - 2015 Jul 23 |
Keywords
- Aliased noise
- Computed tomography (CT)
- Modulation transfer function (MTF)
- Noise power spectrum (NPS)
- Signal-to-noise ratio (SNR)
- Two-alternative forced choice (2-AFC)
ASJC Scopus subject areas
- Radiation
- Physical Therapy, Sports Therapy and Rehabilitation
- Radiology Nuclear Medicine and imaging