Hybrid Randoms Variance Reduction in PET Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current positron emission tomography (PET) systems face significant noise due to random coincidences, which increase data variance and degrade image quality, requiring effective methods for noise reduction that are easy to implement and computationally efficient.
Innovation Solution
A method and system that estimate and correct for randoms variance by averaging fan sums from the delayed sinogram and singles rates from the sinogram header, using a hybrid approach that leverages both available data for precise noise reduction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-generated harmful factors
If random coincidence correction is performed by subtracting delayed coincidences from prompt coincidences, then random coincidences are removed from the data, but noise variance is added to the corrected measured events
Solution Approach 1:
The patent changes the parameter used for random coincidence estimation from the traditional delayed coincidence window approach to a hybrid method using fan sum data and singles rates. This parameter change allows for a more accurate estimation of random coincidences, reducing the noise variance added during correction while maintaining the ability to remove random events from the data.
Solution Approach 2:
The patent replaces the conventional mechanical approach of using delayed coincidence windows with a computational method that uses fan sum data and singles rates from the sinogram header. This substitution enables more precise random coincidence estimation and reduces the impact of noise variance in the corrected data.
2Ease of operation
If conventional random coincidence estimation methods are used, then the correction can be applied, but the noise in the estimated random coincidences increases data variance
Solution Approach 1:
The patent introduces fan sum data and singles rates as intermediary elements to estimate random coincidences more accurately. These intermediaries allow for a better separation of signal and noise, enabling the correction to be applied while minimizing the increase in data variance.
Solution Approach 2:
The patent uses feedback from the sinogram header (singles rates) and fan sum data to continuously improve the estimation of random coincidences. This feedback mechanism allows for optimized correction that reduces data variance while maintaining correction effectiveness.
3Measurement precision
If more accurate randoms variance reduction methods are implemented, then image quality improves, but computational complexity and implementation difficulty increase
Solution Approach 1:
The patent segments the random coincidence correction process into distinct steps: estimating fan sums from delayed sinogram data, obtaining singles rates from the sinogram header, and combining these to calculate corrected prompt coincidences. This segmentation makes the complex correction process more manageable and easier to implement while maintaining image quality improvements.
Solution Approach 2:
The patent creates a universal correction method that can be applied to various PET scanning scenarios using the same fan sum and singles rate approach. This multi-functional method improves image quality across different applications without requiring complex scenario-specific implementations.
Data Source
AI summary
A method for reducing randoms variance in a Positron Emission Tomograph (PET) or Positron Emission Tomograph combined with another Medical Imaging device is disclosed. An average of an element of the randoms event (delayeds) sinogram may be estimated by dividing fan sums in delayeds sinogram by singles rates taken from headers of the delayeds sinogram.