Abstract
Dynamic PET image reconstruction is a challenging problem because of the ill-conditioned nature of PET and the lowcounting statistics resulted from short time-frames in dynamic imaging. The kernel method for image reconstruction has been developed to improve image reconstruction of low-count PET data by incorporating prior information derived from high-count composite data. In contrast to most of the existing regularization-based methods, the kernel method embeds image prior information in the forward projection model and does not require an explicit regularization term in the reconstruction formula. Inspired by the existing highly constrained back-projection (HYPR) algorithm for dynamic PET image denoising, we propose in this work a new type of kernel that is simpler to implement and further improves the kernel-based dynamic PET image reconstruction. Our evaluation study using a physical phantom scan with synthetic FDG tracer kinetics has demonstrated that the new HYPR kernel-based reconstruction can achieve a better region-of-interest (ROI) bias versus standard deviation trade-off for dynamic PET parametric imaging than the post-reconstruction HYPR denoising method and the previously used nonlocal-means kernel.
| Original language | English (US) |
|---|---|
| Title of host publication | Medical Imaging 2017 |
| Subtitle of host publication | Physics of Medical Imaging |
| Publisher | SPIE |
| Volume | 10132 |
| ISBN (Electronic) | 9781510607095 |
| DOIs | |
| State | Published - 2017 |
| Event | Medical Imaging 2017: Physics of Medical Imaging - Orlando, United States Duration: Feb 13 2017 → Feb 16 2017 |
Other
| Other | Medical Imaging 2017: Physics of Medical Imaging |
|---|---|
| Country | United States |
| City | Orlando |
| Period | 2/13/17 → 2/16/17 |
Fingerprint
ASJC Scopus subject areas
- Atomic and Molecular Physics, and Optics
- Electronic, Optical and Magnetic Materials
- Biomaterials
- Radiology Nuclear Medicine and imaging
Cite this
Dynamic PET Image reconstruction for parametric imaging using the HYPR kernel method. / Spencer, Benjamin; Qi, Jinyi; Badawi, Ramsey D; Wang, Guobao.
Medical Imaging 2017: Physics of Medical Imaging. Vol. 10132 SPIE, 2017. 101324W.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Dynamic PET Image reconstruction for parametric imaging using the HYPR kernel method
AU - Spencer, Benjamin
AU - Qi, Jinyi
AU - Badawi, Ramsey D
AU - Wang, Guobao
PY - 2017
Y1 - 2017
N2 - Dynamic PET image reconstruction is a challenging problem because of the ill-conditioned nature of PET and the lowcounting statistics resulted from short time-frames in dynamic imaging. The kernel method for image reconstruction has been developed to improve image reconstruction of low-count PET data by incorporating prior information derived from high-count composite data. In contrast to most of the existing regularization-based methods, the kernel method embeds image prior information in the forward projection model and does not require an explicit regularization term in the reconstruction formula. Inspired by the existing highly constrained back-projection (HYPR) algorithm for dynamic PET image denoising, we propose in this work a new type of kernel that is simpler to implement and further improves the kernel-based dynamic PET image reconstruction. Our evaluation study using a physical phantom scan with synthetic FDG tracer kinetics has demonstrated that the new HYPR kernel-based reconstruction can achieve a better region-of-interest (ROI) bias versus standard deviation trade-off for dynamic PET parametric imaging than the post-reconstruction HYPR denoising method and the previously used nonlocal-means kernel.
AB - Dynamic PET image reconstruction is a challenging problem because of the ill-conditioned nature of PET and the lowcounting statistics resulted from short time-frames in dynamic imaging. The kernel method for image reconstruction has been developed to improve image reconstruction of low-count PET data by incorporating prior information derived from high-count composite data. In contrast to most of the existing regularization-based methods, the kernel method embeds image prior information in the forward projection model and does not require an explicit regularization term in the reconstruction formula. Inspired by the existing highly constrained back-projection (HYPR) algorithm for dynamic PET image denoising, we propose in this work a new type of kernel that is simpler to implement and further improves the kernel-based dynamic PET image reconstruction. Our evaluation study using a physical phantom scan with synthetic FDG tracer kinetics has demonstrated that the new HYPR kernel-based reconstruction can achieve a better region-of-interest (ROI) bias versus standard deviation trade-off for dynamic PET parametric imaging than the post-reconstruction HYPR denoising method and the previously used nonlocal-means kernel.
UR - http://www.scopus.com/inward/record.url?scp=85020393643&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85020393643&partnerID=8YFLogxK
U2 - 10.1117/12.2254497
DO - 10.1117/12.2254497
M3 - Conference contribution
AN - SCOPUS:85020393643
VL - 10132
BT - Medical Imaging 2017
PB - SPIE
ER -