标题： 能量阈值偏差计算器：一种基于物理模型的光子计数CT自适应校正方案 显示英文标题

标题： Energy-Threshold Bias Calculator: A Physics-Model Based Adaptive Correction Scheme for Photon-Counting CT

摘要： 基于光子计数探测器的计算机断层扫描（PCCT）近年来取得了很大进展。 然而，光谱不一致性仍然是PCCT的一个严重挑战，可能会直接引入明显的伪影和严重的不准确性。 本工作尝试通过在一种新颖、统一且两项因子分解框架中对光谱不一致性进行建模来克服这一挑战，其中包括一个与能量阈值无关的光谱偏斜项，以及一个能量阈值偏差的分析表征项。 为了解决光谱不一致性，这里推导出了一种称为能量阈值偏差计算器（ETB-Cal）的两步分解算法，在该算法中，光谱偏斜项在相对较低的能量阈值下被大致确定，而仅需计算能量阈值偏差，随着能量阈值的变化。 在校准阶段计算出这两个项后，它们将被纳入我们的光谱模型中，以生成光谱校正向量以及如需要的材料分解向量，用于PCCT投影数据。 为了验证我们的方法，在桌面PCCT系统上进行了数值模拟和物理实验。 初步结果表明，光谱不一致性可以显著减少，例如，Gammex多能幻影的非均匀性定量指标从26.27 HU降至5.80 HU，京都头颅幻影的非均匀性定量指标从27.88 HU降至3.16 HU。 显著的改进一致表明，所提出的基于新物理模型的校正方案在实际应用中具有很大的潜力，其计算效率高，校准方便，通用性强，并且基本上避免了在能量阈值校准中使用X射线荧光材料。 显示英文摘要

摘要： Photon-counting detector based computed tomography (PCCT) has greatly advanced in recent years. However, the spectral inconsistency is still a serious challenge for PCCT that could directly introduce obvious artifacts and severe inaccuracies. This work attempts to overcome the challenge by modeling the spectral inconsistency in a novel, unified, and two-term factorized framework, with a spectral skew term independent of the energy threshold, and an energy-threshold bias analytical characterization term. To solve the spectral inconsistency, a two-step decomposition algorithm called energy-threshold bias calculator (ETB-Cal) is derived here, in which the spectral skew term is grossly determined at a relatively low energy threshold and only the energy-threshold bias is needed to be calculated as the energy threshold changes. After the two terms being computed out in calibration stage, they will be incorporated into our spectral model to generate the spectral correction vectors as well as the material decomposition vectors if needed, for PCCT projection data. To validate our method, both numerical simulations physics experiments were carried out on a tabletop PCCT system. Preliminary results showed that the spectral inconsistency can be significantly reduced, for example, with an non-uniformity quantitative indicators decreasing from 26.27 to 5.80 HU for Gammex multi-energy phantom and from 27.88 to 3.16 HU for kyoto head phantom. The considerable improvements consistently demonstrate a great potential of the proposed novel physics-model based correction scheme in practical applications, as computationally efficient, calibration-wise convenient with high degree of generality, and substantially avoiding the use of X-ray florescence material in the energy-threshold calibration.