本研究旨在同步改善飛行時間正子掃描系統(Time-of-Flight PET)之放射活性及衰減估算。近年來PET硬體時間解析度的提高,使PET系統得以解決飛行時間的資訊,能夠減少雜訊和假影,從而提高影像的質量。而透過疊代式的演算法,TOF-PET已經可以進一步延伸到同步估算光子活動以及衰減的分佈圖;在混合的正子掃描/磁振系統中,磁振影像(MRI)可以提供良好的軟組織對比度和解析度,並且也可以用於衰減校正。 在這項研究中,我們將數據保真項、總變異和磁振影像衍生之衰減事前資訊表達成一個複合的凸優化問題。根據我們透過電腦程式設計演算法及模擬的結果,與使用最大似然放射活性與衰減估算演算法(MLAA)獲得的放射活性和衰減估計進行比較,加入磁振事前資訊與總變異約束項的MR-MLAA-TV演算法的應用可以有效提高影像邊緣特徵,並降低雜訊的出現。
The aim of this study was to improve the radio-activity and attenuation estimates of time-of-flight PET. In recent years, as the time resolution of PET hardware improved, the time-of flight (TOF) information can be fully utilized, which can reduce noise and artifacts of the PET reconstruction image, thereby improving the quality of the image. TOF-PET can be further extended to simultaneously estimate the photon activity and the attenuation distribution. In hybrid PET/MR systems, the magnetic resonance image (MRI) can offer excellent soft tissue contrast and resolution and also be used for attenuation correction. In this study, we express the data fidelity term, the total variation term and the MR attenuation prior information into a composite convex optimization problem. According to the results of the computer program design algorithm and simulation, we compare the activity and attenuation estimation results of obtained by MLAA algorithm with MR-MLAA-TV algorithm. The application of the MR-MLAA-TV algorithm with the MR prior information and the total variation constraint can effectively improve the edge feature of the image and reduce the occurrence of noise.