物理
探测器
蒙特卡罗方法
光子
图像分辨率
康普顿散射
多普勒效应
概率逻辑
多普勒展宽
能量(信号处理)
光学
角度分辨率(图形绘制)
算法
计算机科学
人工智能
谱线
数学
统计
量子力学
组合数学
天文
作者
Chuanpeng Wu,Siyuan Zhang,Liang Li
标识
DOI:10.1088/1361-6560/ac73d2
摘要
Abstract Objective . The Compton cameras have been researched for medical applications and radioactive material detection. It is challenging for the Compton camera to realize high-resolution reconstruction when the incident photon energy is below 200 keV. However, multiple kinds of nuclear medical radionuclides are in this energy range, such as 201 Tl, 67 Ga, 99m Tc, and 123 I. In this work, we propose an improved probabilistic model with correction of detector energy resolution, detector spatial resolution, and Doppler broadening effect. The proposed model is used for numerical calculation of the system matrix in the list-mode maximum likelihood expectation maximization (LM-MLEM) algorithm. Approach. The model can improve the imaging resolution of LM-MLEM reconstruction by taking Doppler broadening effect into account. It performs well, especially in the following situations: low-energy photon incidence below 200 keV or (and) small distance between scattering and absorbing positions. Main results. Firstly, three main factors that affect the angular resolution of the Compton camera are theoretically analyzed and quantitatively calculated. The results of the analysis indicate the necessity of including the Doppler broadening effect in the model. Secondly, the details and derivation of the proposed probabilistic model are described. Thirdly, both Monte Carlo (MC) simulations and experiments are carried out to verify the performance of the proposed algorithm. The simulations focus on the low-energy reconstruction in which 201 Tl (70 keV) and 99m Tc (141 keV) are simulated. And the experiments are based on a single-layer Compton camera composed of a Timepix3 detector. Significance. The results of the simulations and the Timepix3-based experiments are presented to verify the effectiveness of the proposed algorithm. The model improves the Compton imaging resolution when the photon energy is below 200 keV.
科研通智能强力驱动
Strongly Powered by AbleSci AI