Speaker
Description
Signal, noise, and their correlations– in space are fundamental for determining image quality. These can be characterized by modulation-transfer function (MTF) and noise-power spectrum (NPS), which are both Fourier metrics. The ratio of the squared MTF and the NPS is termed noise-equivalent quanta (NEQ) or detective quantum efficiency (DQE), commonly used to describe the performance of an imaging system.
The cascaded-systems analysis (CSA) approach has been used to parameterize the signal and noise characteristics of 2D radiographic systems in terms of operational and design parameters, such as incident photon fluence, quantum efficiency, Swank factor, electronic noise, and pixel fill factor [1]. These parameters, explicitly expressed in the CSA formula, are implicitly related to the factors like the dose used for imaging, detector thickness, dark current, amplifier noise, and more. Consequently, CSA can identify parameters that degrade system performance, offering pathways for system improvement or optimization.
CSA can also be applied to computed tomography (CT). The CSA for cone-beam CT (CBCT) using flat-panel detectors requires additional filtering and backprojection processes. While the mathematical 3D image reconstruction process is deterministic, it has been shown to be irreversible due to sampling processes [2]. As a result, acquisition and reconstruction parameters, such as the number of projection views and reconstruction filters, significantly influence the 3D noise characteristics (e.g., NEQ or DQE). Therefore, developing a theoretical model to describe the 3D imaging chain in CBCT is essential for optimizing system parameters based on specific applications.
In this study, we apply CSA to CBCT using robotic arms, instead of the conventional circular gantry. In the filtered backprojection algorithm, one-directional filtering (i.e., filtering parallel to the scan direction) leads to a 3D NPS that is asymmetric between the axial and longitudinal directions, as shown in Fig. 1. The backprojection of projections obtained from arbitrary scan trajectories may further degrade the 3D NPS, an effect that is explored in this study. We examine the noise properties of robot CT with arbitrary scan paths for various operations and design parameters using the CSA model, comparing the results with measurements. The developed model can be used to determine the optimal scan path for robot CT.
Reference
[1] H. K. Kim, Ch. 7 in Radiation Detection Systems, Boca Raton, FL, CRC Press (2021) 163-224
[2] D. J. Tward and J. H. Siewerdsen, Med. Phys. 35(12), 5510-5529 (2008)
Acknowledgement
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2021R1A2C1010161 and RS-2024-00340520).
| Workshop topics | Applications |
|---|
