PET Event Data Resampling for Noise Reduction and Image Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for improving PET image quality, such as bootstrap resampling, face challenges in clinical application due to extensive data processing requirements and noise amplification, particularly in gated PET data and low count rates, which degrade image quality and disease detection capabilities.
Innovation Solution
An apparatus and method for resampling event data from PET scanners by analyzing list data, replicating and reconstructing packet data of specific event types, and adjusting gray codes to increase data quantity and improve noise and statistical characteristics, thereby enhancing image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If bootstrap resampling is used to improve PET image quality, then noise reduction is achieved, but extensive data processing requirements increase device complexity and processing time
Solution Approach 1:
The patent segments the list data into individual event records with specific identifiers (coincidence events, single events, random events). By separating and selectively resampling only the relevant coincidence events rather than processing all data types, the method reduces overall data processing complexity while maintaining noise reduction effectiveness.
Solution Approach 2:
The patent applies partial resampling by selectively duplicating only the necessary event records (coincidence events) rather than resampling the entire dataset. This partial action approach achieves the statistical improvement needed for noise reduction while avoiding the computational burden of processing excessive data.
2Object-affected harmful factors
If filtering techniques are used to remove noise from PET images, then noise is reduced, but image resolution is degraded
Solution Approach 1:
The patent performs resampling of event data before image reconstruction rather than filtering after reconstruction. This preliminary action increases the statistical quality of the input data, allowing the reconstruction algorithm to produce high-resolution images with reduced noise inherently, rather than requiring post-processing filters that would degrade resolution.
3Reliability
If data resampling is performed to increase count rate, then matching rate between multimodal images is improved, but data processing time is increased
Solution Approach 1:
The patent applies resampling selectively to specific event types (coincidence events) that are most critical for image quality and matching rate, rather than uniformly processing all event data. This localized approach to data quality improvement achieves the necessary count rate increase for better multimodal matching while minimizing unnecessary processing time expenditure on less critical data.
Data Source
AI summary
Provided are apparatuses and methods of resampling event data for quantitative improvement of a PET image, which acquire a quantitatively-improved new PET image by analyzing a storage format of list data prior to conversion into the PET image and nonparametrically resampling event data from the list data based on the analysis results to improve noise and statistical characteristics thereof. The quantitatively improved new PET image may be acquired by analyzing a storage format of list data constituting the PET image and nonparametrically resampling event data to improve noise and statistical characteristics thereof.