Multi-Energy CT Noise Reduction via Spatiospectral Redundancy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional CT imaging faces challenges with high image noise and partial volume effects, which compromise diagnostic quality and image resolution, especially in multi-energy CT applications where noise amplification and slice thickness trade-offs hinder the differentiation of materials and anatomical features.
Innovation Solution
A system and method that exploit redundant CT image information to reduce noise and partial volume effects by using a computer system to process CT data with an objective function incorporating total variation and spatiospectral redundancy, promoting sparsity and minimizing noise while retaining structural details.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multi-energy CT is used to differentiate materials, then material differentiation capability is improved, but image noise increases
Solution Approach 1:
The patent combines multiple energy-specific images into a synthesized image that integrates information from different energy levels. This merging process allows material differentiation capabilities of multi-energy CT while reducing the noise penalty through complementary information fusion across energy bins.
Solution Approach 2:
The patent introduces an intermediate synthesized image as a mediator between raw multi-energy data and final material-specific images. This intermediate representation captures complementary information from multiple energy bins and serves as a noise-reduced foundation for subsequent material differentiation tasks.
2Adaptability or versatility
If radiation dose is divided into multiple energy bins, then multi-energy CT capability is improved, but image noise within each energy-specific image increases
Solution Approach 1:
The patent merges multiple noisy energy-specific images into a single synthesized image that benefits from the combined signal-to-noise ratio. By integrating information across all energy bins, the synthesized image achieves lower noise levels than any individual energy-specific image while preserving multi-energy CT functionality.
3Measurement precision
If material decomposition is performed to identify and quantify materials, then material identification capability is improved, but image noise is amplified
Solution Approach 1:
The patent performs preliminary synthesis of a low-noise intermediate image from multiple energy bins before conducting material decomposition. This preliminary action creates a cleaner input for the decomposition algorithm, reducing the amplification of noise that typically occurs during material identification and quantification processes.
4Manufacturing precision
If thin slice thickness is used, then partial volume effect is reduced, but image noise increases
Solution Approach 1:
The patent combines information from multiple energy bins to create a synthesized image that maintains thin slice thickness for reduced partial volume effects. The merging process accumulates photon statistics across energy levels, providing sufficient signal-to-noise ratio even with thin slice geometry that would normally be too noisy.
Data Source
AI summary
A system and method is provided for high fidelity multi-energy CT processing. This system and method exploits prior knowledge, where prior knowledge may include redundant information existing in the CT images, such as spatial redundancy between a thick slice and a thin slice encompassed by or close to the thick slice, or the spatiospectral redundancy between the image output of multi-energy CT processing and the source multi-energy CT images. The system and method retains structural details, spatial resolution, spectral fidelity, and noise texture while achieving noise reduction. The method reduces image noise and increases the contrast-to-noise ratio in processed images, while simultaneously maintaining image details and natural appearance of the image to enhance detectability and facilitate reader acceptance.