Mean Randoms Estimation in PET Scanners Using Dynamic Frame Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional PET scanners struggle to adequately account for varying singles rates over time, particularly in scenarios like Continuous Bed Motion (CBM) scans and stationary scans with rapidly changing tracer distributions, leading to noise sensitivity and suboptimal image reconstruction.
Innovation Solution
The proposed solution involves using a randoms smoothing model (rij=∫si(t)sj(t)dt) to estimate mean randoms, which provides a more accurate representation of singles rates over time. This model allows for improved PET image reconstruction without the need for additional heuristic methods that may introduce data correction errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional delay logic is used to estimate randoms, then the estimation process is simple, but the accuracy deteriorates when singles rate varies in time
Solution Approach 1:
The patent divides the acquisition time into multiple frames and calculates delay coincidences separately for each frame. This dynamic approach allows the estimation to adapt to time-varying singles rates, resolving the contradiction by making the estimation process frame-dependent rather than using a single static delay coincidence measurement for the entire acquisition.
Solution Approach 2:
The patent segments the acquisition time into discrete frames and performs separate delay coincidence calculations for each frame. This segmentation enables accurate tracking of singles rate variations over time while maintaining a manageable computational structure, thus improving accuracy without excessive complexity.
2Ease of manufacture
If plane-by-plane scaling is applied to mean randoms, then the processing is straightforward, but noise sensitivity increases particularly for oblique segments
Solution Approach 1:
The patent applies different scaling factors to different LORs based on their specific characteristics. Instead of uniform plane-by-plane scaling, each LOR receives a customized scaling factor derived from its delay coincidences and singles rates, which reduces noise sensitivity particularly for oblique segments while maintaining ease of implementation through LOR-specific corrections.
3Productivity
If delay coincidence counts are used for rescaling, then the method is computationally efficient, but accuracy deteriorates in CBM scans and scans with rapidly changing tracer distribution
Solution Approach 1:
The patent dynamically updates delay coincidence measurements for each frame in CBM scans and scans with rapidly changing tracer distribution. This frame-by-frame dynamic measurement maintains computational efficiency while accurately capturing time-varying conditions, resolving the contradiction between efficiency and accuracy in dynamic scanning scenarios.
Solution Approach 2:
The patent performs delay coincidence measurements and singles rate calculations for each frame before the actual randoms correction is applied. This preliminary action ensures that the most current and accurate singles rate information is available for scaling, improving accuracy while maintaining computational efficiency through pre-calculation.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The improved estimation of mean randoms using the proposed model leads to enhanced PET image quality, particularly in scenarios with varying singles rates, resulting in reduced noise and improved signal-to-noise ratio.
Implementation Method 1
Radioactive decay of the tracer generates positrons which eventually encounter electrons and are annihilated thereby
Implementation Method 2
positrons which eventually encounter electrons and are annihilated thereby. Annihilation produces two photons
Implementation Method 3
Each crystal element comprises a scintillator
Data Source
AI summary
Systems and methods to estimate mean randoms include acquisition of list mode data describing true coincidences and delay coincidences detected during a scan of an object, determination of a plurality of time periods of the scan based on a distance moved by a bed supporting the object during each of the plurality of time periods, determination, for each crystal and for each of time period, of delay coincidences including the crystal based on the list mode data, determination, for each crystal, of a singles rate associated with each time period based on the delay coincidences determined for the crystal over the time period, determination, for each time period, of estimated mean randoms for each crystal pair based on the singles rate associated with the time period, and reconstruction of an image of the object based on the estimated mean randoms for each time period and the detected true coincidences.