WEKO3
アイテム
Development of Scatter Correction with Image-domain Interpolation for TOF Helmet-type PET
https://repo.qst.go.jp/records/83757
https://repo.qst.go.jp/records/83757f42377b0-9be8-4b0a-9a43-f1fa2a1387dd
Item type | 会議発表用資料 / Presentation(1) | |||||
---|---|---|---|---|---|---|
公開日 | 2021-10-06 | |||||
タイトル | ||||||
タイトル | Development of Scatter Correction with Image-domain Interpolation for TOF Helmet-type PET | |||||
言語 | ||||||
言語 | eng | |||||
資源タイプ | ||||||
資源タイプ識別子 | http://purl.org/coar/resource_type/c_c94f | |||||
資源タイプ | conference object | |||||
アクセス権 | ||||||
アクセス権 | metadata only access | |||||
アクセス権URI | http://purl.org/coar/access_right/c_14cb | |||||
著者 |
Hideaki, Tashima
× Hideaki, Tashima× Go, Akamatsu× Taichi, Yamashita× Taiga, Yamaya× Hideaki, Tashima× Go, Akamatsu× Taichi, Yamashita× Taiga, Yamaya |
|||||
抄録 | ||||||
内容記述タイプ | Abstract | |||||
内容記述 | Scatter correction is essential for quantitative PET imaging. The current standard method corrects the scatter effect on each measured data bin in the projection domain with a scatter component estimated by model-based scatter simulation. However, the scatter components are not calculated for all the data bins due to the computational burden. Instead, coarsely sampled data bins are interpolated for each measured data bin. In our helmet-type PET, the interpolation for the hemispherical arrangement of block detectors is complex compared with conventional cylindrical PET scanners with large gantry bore diameters. Therefore, we propose an image domain scatter correction method to avoid interpolating data in the projection domain to improve correction accuracy. We modeled an uncorrected image as the sum of a radioactivity image and a scatter image. The proposed method alternatively updates the uncorrected radioactivity image and the scatter image. To update the scatter image, we first reconstructed a coarsely sampled scatter image from coarsely sampled scatter components estimated by model-based scatter simulation with a radioactivity image generated by subtracting the latest scatter image from the latest uncorrected image. Then the coarsely sampled scatter image was up-sampled to a fully sampled scatter image. Finally, we obtained a scatter corrected image by subtracting the latest scatter image from the latest uncorrected radioactivity image. We implemented the proposed method with the list-mode OSEM and the single scatter simulation for the time-of-flight helmet-type PET prototype. An experiment with the Hoffman brain phantom showed that the proposed method improved the gray matter to white matter contrast from 45% to 51% compared with the conventional correction method with projection domain interpolation. | |||||
会議概要(会議名, 開催地, 会期, 主催者等) | ||||||
内容記述タイプ | Other | |||||
内容記述 | IEEE NSS MIC 2021 | |||||
発表年月日 | ||||||
日付 | 2021-10-20 | |||||
日付タイプ | Issued |