PCA Calibration for Multiplexed PET Detector Modules
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current PET imaging systems require large amounts of non-multiplexed data for calibration, which increases cost and complexity, and existing multiplexing methods do not effectively preserve information for scatter rejection and depth estimation.
Innovation Solution
A method using principal component analysis (PCA) to simulate and generate weighted multiplexed calibration data directly from the detector module, reducing the need for extensive data collection and processing, and enabling accurate depth-of-interaction estimation and scatter filtering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional calibration methods are used to obtain accurate photon positioning data, then measurement precision is improved, but device complexity and data processing requirements increase significantly
Solution Approach 1:
The patent extracts and removes scattered photons and noise from the calibration dataset using statistical filtering methods. By identifying and eliminating contaminated data points based on energy deposition patterns and spatial distribution, the method obtains clean calibration data without requiring complex hardware filtering systems, thus maintaining measurement precision while reducing device complexity.
Solution Approach 2:
The patent uses simulated photon interactions to create a virtual model of the detector response. By generating synthetic calibration data through Monte Carlo simulation that replicates real photon detection patterns, the method avoids the need for extensive physical calibration measurements and complex calibration hardware, reducing device complexity while maintaining positioning accuracy.
2Productivity
If scattered photons are not filtered from calibration data, then data acquisition is simplified, but measurement precision deteriorates due to contaminated calibration datasets
Solution Approach 1:
The patent implements an iterative feedback-based filtering process where the calibration data is repeatedly analyzed and refined. Scatter rejection criteria are applied based on feedback from energy deposition patterns and spatial distribution analysis, continuously improving data quality while maintaining acquisition efficiency. This automated feedback loop eliminates manual data cleaning while ensuring high lookup table accuracy.
Solution Approach 2:
The patent replaces physical scatter rejection hardware with computational filtering methods. Instead of using mechanical collimators or physical shields to block scattered photons during calibration, the method uses software-based statistical filtering and simulation to identify and remove scattered event data, maintaining measurement precision while simplifying the calibration system.
3Device complexity
If multiplexed detector output is used to reduce data dimensions, then device complexity is reduced, but information loss occurs for depth-of-interaction estimation
Solution Approach 1:
The patent performs preliminary depth-of-interaction estimation using the full multiplexed detector output before any data reduction or lookup table generation. By calculating DOI values in advance from the complete signal information, the method preserves depth information that would otherwise be lost during data processing, enabling accurate 3D positioning while still benefiting from multiplexed readout simplicity.
Solution Approach 2:
The patent transforms the multiplexed detector signals into a different dimensional representation that preserves depth information. By using the temporal and spatial characteristics of the multiplexed signals in combination with simulation-based models, the method extracts DOI information along a new dimension without requiring the full complexity of individual photodetector channel data.
Data Source
AI summary
A method for calibrating a nuclear medicine tomography detector module using principal component analysis is based on the idea that calibration beam data lies on a one-dimensional path within the higher dimensional dataspace of output data. The module includes a weighted multiplexing circuit that generates a small number of multiplexed signals for each photon event. Calibration data for the module is generated and analyzed using several iterations of principal component analyses, to filter scattering events, noise, and other spurious signals. The direction of depth-of-interaction information has been found in the high-dimensional dataspace to be indicated by the primary principal component of the calibration data. The primary principal components, principal components from filtered datasets, intermediate thresholds, and DOI or inner product values are recorded for calibrating the module.