Iterative Image Reconstruction Dynamic Noise Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current iterative image reconstruction technologies in radiological imaging face challenges in effectively suppressing noise-induced artifacts while preserving real features, often leading to erroneous radiology findings due to insufficient or excessive noise suppression.
Innovation Solution
The approach involves adjusting the edge preservation threshold in iterative reconstruction processes, leveraging the difference in process flow evolution between real features and noise-induced artifact features, and employing local detection and suppression of artifact features by replacing pixel or voxel values with earlier image updates or neighboring values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If stronger noise suppression is applied during iterative reconstruction, then noise-induced artifact features are reduced, but real physical features with low contrast may be suppressed leading to erroneous radiology findings
Solution Approach 1:
The patent applies dynamics by making the edge preservation parameter K variable rather than fixed. Specifically, K is adjusted based on the iteration number and the local image features at each iteration stage. This dynamic adjustment allows the reconstruction algorithm to preserve edges effectively during early iterations when real features are being established, while allowing stronger noise suppression in later iterations when the image converges, thus resolving the contradiction between noise suppression and feature preservation
Solution Approach 2:
The patent changes the parameter K (edge preservation strength) during the iterative reconstruction process. The parameter is modified based on iteration count and local image characteristics, transitioning from higher values that preserve edges to lower values that suppress noise. This parameter change strategy enables the system to adaptively balance between preserving real features and suppressing noise-induced artifacts at different stages of reconstruction
2Reliability
If weaker noise suppression is applied during iterative reconstruction, then real physical features are preserved, but noise-induced artifact features may obscure these features leading to erroneous radiology findings
Solution Approach 1:
The patent implements periodic action by applying different levels of noise suppression at different iteration periods. During early iterations, the algorithm uses parameters that prioritize feature preservation with weaker noise suppression. In later iterations, it transitions to stronger noise suppression as the image converges. This periodic adjustment of suppression strength throughout the reconstruction process resolves the contradiction by applying the right level of suppression at the right time
Solution Approach 2:
The patent maintains continuous useful action by progressively adjusting the noise suppression strength throughout the iterative process. Rather than applying a fixed level of suppression, the algorithm continuously adapts the suppression parameter K across iterations, ensuring that both feature preservation and noise suppression are continuously optimized. This continuous adaptation allows the system to maintain reliability while gradually reducing artifacts
3Ease of manufacture
If a fixed threshold is used for edge preservation in iterative reconstruction, then the reconstruction process is simple to implement, but it cannot adapt to varying noise levels and feature contrasts across different regions and iterations
Solution Approach 1:
The patent transforms the static threshold approach into a dynamic one by making the edge preservation parameter K dependent on iteration number and local image features. This dynamic parameter adapts to varying noise levels and feature contrasts automatically during the reconstruction process, resolving the contradiction between implementation simplicity and adaptability while maintaining reasonable computational complexity
Solution Approach 2:
The patent applies local quality by allowing the edge preservation parameter K to vary across different spatial regions and iteration stages. Instead of using a uniform threshold, the algorithm adjusts K based on local image characteristics such as gradient magnitude and noise levels in different regions. This local adaptation enables the reconstruction to handle varying noise and feature conditions effectively while building upon the simple fixed-threshold framework
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Iterative reconstruction (20) of imaging data is performed to generate a sequence of update images (22) terminating at a reconstructed image. During the iterative reconstruction, at least one of an update image and a parameter of the iterative reconstruction is adjusted using an adjustment process separate from the iterative reconstruction. In some embodiments using an edge-preserving regularization prior (26), the adjustment process (30) adjusts an edge preservation threshold to reduce gradient steepness above which edge preservation applies for later iterations compared with earlier iterations. In some embodiments, the adjustment process includes determining (36, 38) for each pixel, voxel, or region of a current update image whether its evolution prior to the current update image (22) satisfies an artifact feature criterion. A local noise suppression operation (40) is performed on the pixel, voxel, or region if the evolution satisfies the artifact feature criterion and is not performed otherwise.