WEKO3
アイテム
Implementation of 3D image reconstruction with a pre-computed system matrix for the jPET-D4.
https://repo.qst.go.jp/records/62152
https://repo.qst.go.jp/records/62152a147eaea-7f7c-4b20-89f0-6641a642c396
| Item type | 会議発表用資料 / Presentation(1) | |||||
|---|---|---|---|---|---|---|
| 公開日 | 2007-07-18 | |||||
| タイトル | ||||||
| タイトル | Implementation of 3D image reconstruction with a pre-computed system matrix for the jPET-D4. | |||||
| 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 |
Yamaya, Taiga
× Yamaya, Taiga× Yoshida, Eiji× Lam, ChihFung× Obi, Takashi× Murayama, Hideo× 山谷 泰賀× 吉田 英治× Ｌａｍ ＣｈｉｈＦｕｎｇ× 小尾 高史× 村山 秀雄 |
|||||
| 抄録 | ||||||
| 内容記述タイプ | Abstract | |||||
| 内容記述 | The jPET-D4 is a novel brain PET scanner which aims to achieve not only high spatial resolution but also high scanner sensitivity by using 4-layer depth-of-interaction (DOI) information. The dimensions of a system matrix for the jPET-D4 become 3.3 billion (coincidence pairs) x 5 million (image elements) when a 25cm diameter FOV is sampled by a (1.5mm)3 voxel. The size of the system matrix is estimated at 117 petabytes (PB) with the accuracy of 8 bytes per element. An on-the-fly calculation is usually used to deal with such a huge system matrix. However we can not avoid extension of the calculation time when we improve the accuracy of system modeling. In this work, we develop an alternative approach based on pre-calculation of the system matrix. The 117PB system matrix is compressed into 24.8GB by (1) eliminating zero elements, (2) applying the DOI compression method and (3) applying rotation symmetry and an axial shift property of the crystal arrangement. Spanning, which degrades axial resolution, is not applied. Histogram-based 3D OS-EM based on geometrical system modeling is implemented on a desktop workstation with 2 dual-core CPUs and 32GB memory installed. After optimizing parallel processing, data access ordering, the number of maximum ring difference and the number of subsets, we can reduce calculation time to around 1 hour. | |||||
| 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 | Fully Three-Dimensional Image Reconstruction Meeting in Radiology and Nuclear Medicine | |||||
| 発表年月日 | ||||||
| 日付 | 2007-07-13 | |||||
| 日付タイプ | Issued | |||||
