Related Experiment Video
Updated: Mar 13, 2026
12:10
Single Particle Electron Microscopy Reconstruction of the Exosome Complex Using the Random Conical Tilt Method
Published on: March 28, 2011
24.0K
Reconstruction of multiple-pinhole micro-SPECT data using origin ensembles.
Morgan C Lyon1, Arkadiusz Sitek2, Scott D Metzler3
1Division of Nuclear Medicine and Molecular Imaging, Department of Radiology, Brigham and Women's Hospital, Boston, Massachusetts 02115 and Department of Radiology, Harvard Medical School, Boston, Massachusetts 02115.
Medical Physics
|October 27, 2016
Summary
The origin ensemble (OE) algorithm reconstructs microSPECT scout images similarly to MLEM but faster and with uncertainty estimates. Accelerated OE imaging enables quicker, real-time analysis for high-resolution scans.
Area of Science:
- Medical Imaging
- Nuclear Medicine
- Biomedical Engineering
Background:
- Developing a dual-resolution microSPECT system for whole-body and high-resolution organ imaging in mice.
- Sequential imaging utilizes two multiple-pinhole tungsten collimator tubes for scout and detailed scans.
- Real-time reconstruction of scout images is desired for efficient positioning of high-resolution scans.
Purpose of the Study:
- Investigate the utility of the origin ensemble (OE) algorithm for online and offline reconstruction of microSPECT scout data.
- Evaluate OE algorithm's ability to provide image uncertainty estimates.
- Assess techniques for accelerating OE reconstruction for potential real-time applications.
Main Methods:
- Calculated system matrices for a 39-pinhole scout collimator.
- Simulated SPECT projections using the MOBY digital mouse phantom across various count levels.
- Compared OE and MLEM reconstruction algorithms, evaluating OE convergence via image entropy and VOI counts.
- Assessed parallelization's impact on OE reconstruction time and performance.
Main Results:
- OE and MLEM reconstructed images showed agreement within 6% for key organs and tissues.
- OE image entropy converged around 2000 iterations; heart counts converged earlier at ~200 iterations.
- Accelerated OE completed 1000 iterations in <9 min (vs. ~79 min for conventional OE), with minor entropy performance loss.
- Combined acceleration techniques reduced reconstruction time without performance degradation.
Conclusions:
- OE algorithm provides quantitative and qualitative results similar to MLEM for microSPECT imaging.
- OE offers valuable image uncertainty estimates, unlike MLEM.
- Parallel computing accelerates OE reconstruction, making it suitable for real-time applications.
- The OE algorithm is effective for multiple-pinhole SPECT data reconstruction and adaptable for real-time use.
