Joint Estimation of Tissue Types and Attenuation Coefficients in CT Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for computed tomography (CT) imaging using photon counting detector-based x-ray CT decouple the estimation of energy-dependent linear attenuation coefficients and tissue types, making it difficult to accurately utilize prior information for regularization, especially at organ boundaries.
Innovation Solution
The Joint Estimation Maximum A Posteriori (JE-MAP) method jointly estimates energy-dependent linear attenuation coefficients and tissue types using latent Markov Random Field calculations, Poisson noise models, and Bayesian estimation, incorporating prior information through a voxel-based coupled Markov random field model and Gaussian mixture model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If sequential material decomposition is used to reconstruct density images first, then estimate linear attenuation coefficients and tissue types, then the processing steps are simplified and decoupled, but the accuracy of tissue type identification and linear attenuation coefficient estimation deteriorates due to inability to use a priori information for regularization
Solution Approach 1:
The patent merges the sequential two-step process (material decomposition followed by linear attenuation coefficient estimation) into a unified joint estimation framework. By combining these previously separate steps, the system can simultaneously optimize both density reconstruction and tissue type identification, allowing a priori information about tissue types to be used for regularization of linear attenuation coefficients, thereby improving accuracy without sacrificing the structured approach to processing.
Solution Approach 2:
The patent incorporates a priori information about tissue types as preliminary knowledge before the estimation process begins. This prior information is used to establish constraints and regularization terms that guide the joint estimation algorithm, enabling more accurate tissue type identification and linear attenuation coefficient estimation by leveraging expected tissue characteristics before analyzing the actual photon count data.
2Device complexity
If simple edge-preserving prior is used for regularization, then the computational complexity is reduced, but the regularization effectiveness deteriorates compared to tissue type-based regularization
Solution Approach 1:
The patent applies different regularization strategies to different spatial locations based on tissue type information. By identifying tissue types locally and applying appropriate regularization constraints specific to each tissue type (e.g., bone vs. soft tissue vs. air), the system achieves more effective regularization than uniform edge-preserving methods while maintaining computational tractability through the structured joint estimation framework.
3Measurement precision
If joint estimation of linear attenuation coefficients and tissue types is performed, then the accuracy of tissue type identification and boundary detection is improved, but the computational complexity and algorithm complexity increase
Solution Approach 1:
The patent transforms the complex joint estimation problem into a more tractable form by changing the parameter representation and optimization approach. The joint estimation of linear attenuation coefficients and tissue types is formulated with specific parameterizations that allow the use of efficient optimization algorithms, reducing the computational burden while maintaining the accuracy benefits of joint estimation.
4Stability of the object's composition
If values of neighboring pixels of linear attenuation coefficients are expected to vary smoothly within same tissue type, then the regularization within tissue types is improved, but the discontinuity at organ boundaries may cause smoothing errors if not properly handled
Solution Approach 1:
The patent applies different regularization strengths and types to different spatial locations based on tissue type classification. Within homogeneous tissue regions, strong smoothing regularization is applied to ensure smooth variation of linear attenuation coefficients. At organ boundaries and tissue interfaces, the regularization is adjusted or reduced to preserve sharp discontinuities, achieving both intra-tissue smoothness and inter-tissue boundary sharpness simultaneously through the joint estimation framework.
Data Source
AI summary
The present invention is directed to a new joint estimation framework employing MAP estimation based on pixel-based latent variables for tissue types. The method combines the geometrical information described by latent MRF, statistical relation between tissue types and P-C coefficients, and Poisson noise models of PCD data, and makes possible the continuous Baysian estimation from detected photon counts. The proposed method has better accuracy and RMSE than the method using FBP and thresholding. The joint estimation framework has the potential to further improve the accuracy by introducing more information about tissues in human body, e.g., the location, size, and number of tissues, or limited variation of neighboring tissues, which will be easily formulated by pixel-based latent variables.