Radiation CT Image Processing Reducing Streak Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for reducing fine streak artifacts in radiation tomographic images, such as smoothing processing, are inadequate as they degrade the sharpness of reconstructed images and fail to fully suppress streak artifacts due to limitations in noise level uniformity across different radiation detection levels.
Innovation Solution
An image processing method that applies suppression processing to high noise level portions with radiation detection levels below a specified threshold and enhancement processing to low noise level portions with detection levels above the threshold, adjusting the degree of processing based on detection levels to promote noise level uniformity and reduce streak artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If smoothing processing is applied to suppress noise in high noise level portions, then noise level uniformity is improved, but sharpness of reconstructed image is degraded
Solution Approach 1:
The patent applies different processing strategies to different regions of scan data based on their noise levels. High noise level portions (low radiation detection levels) receive smoothing processing to reduce noise, while low noise level portions (high radiation detection levels) receive enhancement processing to maintain or improve sharpness. This local differentiation resolves the contradiction by avoiding uniform smoothing that would degrade overall image sharpness.
Solution Approach 2:
Instead of applying smoothing uniformly to all data, the patent inverts the conventional approach by applying enhancement processing to low noise level portions. This inversion allows the system to compensate for the blurring effect of smoothing in high noise regions while maintaining sharpness in low noise regions, thereby resolving the trade-off between noise suppression and image sharpness.
2Object-generated harmful factors
If smoothing processing is applied to reduce noise level, then fine streak artifacts are reduced, but the degree of smoothing is limited by side effects
Solution Approach 1:
The patent applies smoothing processing selectively to high noise level portions where fine streak artifacts are most prominent, while applying enhancement processing to low noise level portions. This localized approach allows aggressive noise suppression in problematic regions without uniformly degrading image sharpness across the entire dataset.
Solution Approach 2:
The patent dynamically adjusts processing parameters based on the radiation detection level of each data portion. The degree of smoothing is increased for low detection level portions (high noise) while enhancement processing is applied to high detection level portions. This parameter adaptation allows optimal artifact reduction without excessive smoothing that would compromise image quality.
Data Source
AI summary
To reduce streak-like artifacts more in a radiation tomographic image: There is provided an image producing apparatus comprising: a processing component configured to, in scan data acquired by a radiation CT scan, apply suppression processing with which noise components are suppressed to a high noise level portion having a radiation detection level lower than a specified threshold, and apply enhancement processing with which noise components are enhanced to a low noise level portion having a radiation detection level equal to or higher than the specified threshold; and a reconstructing component configured to reconstruct an image based on the scan data subjected to the processing by the processing component.