X-ray CT Scanner Edge Sharpening for Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
X-ray CT systems face challenges in reducing streak artifacts while maintaining the sharpness of edges in images, particularly when using low-dose radiation, as existing methods often result in blurring at high-contrast structures.
Innovation Solution
An X-ray CT system that includes a raw data acquisition part, an image reconstructor, and a sharpening processor, which applies a smoothing process to raw data from multiple views and then applies more intense sharpening to pixels in the central part of the subject relative to the peripheral part, using a subject model to identify and sharpen edges within the image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a smoothing filter is applied to raw data to reduce streak artifacts, then streak artifacts are reduced, but edge sharpness deteriorates causing blurring at high-contrast structures
Solution Approach 1:
The patent applies different processing strategies to different regions of the image: smooth regions receive stronger smoothing to reduce streak artifacts, while edge regions are preserved with minimal smoothing to maintain sharpness. This is achieved by detecting edge portions and selectively applying smoothing only to non-edge regions during iterative reconstruction.
Solution Approach 2:
The smoothing filter strength is dynamically adjusted based on the local characteristics of the image data. The system adaptively determines the degree of smoothing application according to whether a region is identified as an edge or a smooth region, rather than applying a fixed smoothing level across the entire image.
2Object-affected harmful factors
If the degree of smoothing is increased to strongly reduce streak artifacts, then streak artifacts are significantly reduced, but edge blurring increases
Solution Approach 1:
The system applies strong smoothing only to smooth regions where streak artifacts are present, while completely preserving edge regions from smoothing. This localized approach allows aggressive artifact reduction in appropriate areas without compromising edge sharpness elsewhere in the image.
Solution Approach 2:
The system uses iterative reconstruction with feedback loops that continuously evaluate the image data to identify edge portions. Based on this feedback, the smoothing application is adjusted in subsequent iterations to protect edges while reducing artifacts in smooth regions.
3Manufacturing precision
If weighted addition of original image and smoothed image is performed to reduce artifacts while preserving edges, then image quality improves, but processing time increases
Solution Approach 1:
The system performs edge detection and region classification during the iterative reconstruction process itself, before final image formation. This preliminary identification of edge versus smooth regions allows the system to apply appropriate processing strategies without requiring separate post-processing steps, thereby reducing overall processing time.
Solution Approach 2:
The patent combines the artifact reduction function and edge preservation function into a single unified iterative reconstruction process, rather than performing them as separate sequential operations. This integration eliminates redundant processing steps and reduces total computation time while achieving both goals simultaneously.
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
This approach allows for the generation of images with reduced streak artifacts and sharpened edges within a short amount of time, effectively addressing the tradeoff between artifact reduction and edge blurring.
Implementation Method 1
raw data obtained by applying X-rays to a subject in imaging space and by detecting a distribution of X-ray strength passing through the subject
Data Source
AI summary
An image is generated where an edge of a structure in a subject is sharpened with reducing streak artifacts, within a short amount of time. X-rays at multiple angles are applied to the subject placed in imaging space, and a distribution of X-rays strength passing through the subject is detected to obtain raw data associated with multiple views. After smoothing the raw data associated with multiple views, image reconstruction is performed to generate a smoothed image of a predetermined region to be imaged in the imaging space. In the smoothed image, more intense sharpening process is applied to pixels in a region of a central part of the subject, than the pixels in a region of a peripheral part of the subject.