Modular CT Material Decomposition with MACE Consensus Equilibrium
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing CT reconstruction methods using photon counting detectors (PCD) fail to simultaneously utilize measured spectral information and advanced prior models for modular material decomposition, limiting the ability to correct for beam hardening and exploit mutual information between energy bins, and are not suitable for Deep Silicon detectors.
Innovation Solution
A Multi-Agent Consensus Equilibrium (MACE) framework is employed to balance a forward model agent and a prior model agent, using a detector proximal map to update pathlength sinograms and incorporate physical and empirical knowledge, allowing independent optimization of these agents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional PCD-CT reconstruction methods are used (reconstructing from each spectral bin separately or creating material sinogram estimates), then the reconstruction process is computationally simpler and more straightforward, but the ability to simultaneously utilize measured spectral information and advanced prior models for modular material decomposition is lost
Solution Approach 1:
The reconstruction process is divided into separate modular components: a forward model agent that handles spectral information and a prior model agent that incorporates advanced prior knowledge. Each agent operates independently on specific aspects of the data, allowing complex functionality to be achieved through composition of simpler, specialized modules rather than a monolithic complex system
Solution Approach 2:
A consensus equilibrium mechanism is introduced as an intermediary that balances and integrates the outputs of the forward model agent and prior model agent. This mediator coordinates between the spectral information processing and prior model application, enabling both to work together without direct complex interaction while maintaining their individual advantages
2Measurement precision
If advanced prior models are incorporated into the reconstruction process, then material decomposition accuracy and noise reduction improve, but the computational complexity and processing time increase
Solution Approach 1:
The prior models are pre-computed and prepared in advance as separate agent components. By pre-processing and organizing the prior knowledge structures before the actual reconstruction occurs, the system avoids the need to compute complex prior model applications during the time-critical reconstruction phase, thus maintaining accuracy while reducing processing time
Solution Approach 2:
The reconstruction system dynamically balances between the forward model agent and prior model agent through the consensus equilibrium mechanism. The integration is performed adaptively, allowing the system to leverage pre-computed prior models efficiently while maintaining the ability to incorporate spectral information, thereby optimizing the trade-off between accuracy and processing speed
3Measurement precision
If photon counting detectors are used to increase contrast-to-noise ratio and spatial resolution, then image quality improves, but the ability to correct for beam hardening and exploit mutual information between energy bins is limited
Solution Approach 1:
The forward model agent is designed to handle multiple functions simultaneously: it processes the spectral information from photon counting detectors, corrects for beam hardening effects, and prepares the data for material decomposition. This multi-functional approach ensures that the high-quality spectral data from PCDs is fully utilized without information loss
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 MACE algorithm achieves a 4.5 times better contrast-to-noise ratio (CNR) boost compared to conventional methods, reducing noise and dosage while maintaining spatial resolution, and is applicable to current data processing chains for photon counting CT scanners.
Implementation Method 1
X-ray computed tomography (CT) based on photon counting detectors (PCD) extends standard CT by counting detected photons in multiple energy bins
Implementation Method 2
the forward model agent can represent a conditional distribution of observed data given an unknown CT image
Data Source
AI summary
One or more systems, devices, computer program products and/or computer-implemented methods of use provided herein relate to modular material decomposition from energy resolving photon counting data. According to an embodiment, a system is provided. The system can further comprise a processor that can execute computer-executable components stored in memory, wherein the computer-executable components can comprise a reconstruction component that can reconstruct one or more material decomposed CT images by balancing a forward model agent and a prior model agent, wherein the forward model agent can represent a conditional distribution of observed data given an unknown CT image, wherein the prior model agent can represent an assumed prior distribution, and wherein the observed data can include measurements from a PCD.