CT Image Contrast Banding Correction via Subvolume Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computed tomography (CT) imaging systems produce contrast banding artifacts due to varying contrast levels across multiple acquisitions, leading to image confusion for physicians and patients.
Innovation Solution
A method involving generating an original image with multiple subvolumes, segmenting the image into structures, applying mask-based corrections, and performing streak correction to normalize contrast levels without altering anatomical structures, using a mask-based local DC correction algorithm to adjust zero-frequency components across subvolumes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple acquisitions are combined to improve image quality, then image detail and coverage are improved, but contrast banding artifacts appear due to varying contrast levels across acquisitions
Solution Approach 1:
The image is divided into multiple subvolumes corresponding to different acquisitions. Each subvolume is processed independently to identify and correct contrast banding artifacts while preserving anatomical structures. This segmentation allows targeted correction of artifacts without affecting the entire image uniformly.
Solution Approach 2:
Different correction strategies are applied to different regions of the image based on local characteristics. Anatomical structures are preserved with their original contrast while regions affected by contrast banding artifacts are selectively corrected. This local quality approach ensures that correction is applied only where needed rather than uniformly across the entire image.
2Stability of the object's composition
If contrast normalization is applied across subvolumes, then contrast consistency is improved, but anatomical structures may be altered or distorted
Solution Approach 1:
Anatomical structures are identified and protected before contrast normalization is applied. By pre-segmenting the image to locate anatomical features, the correction algorithm can exclude these regions from normalization or apply adjusted correction factors, ensuring that contrast consistency is improved without distorting critical anatomical structures.
Solution Approach 2:
A protection mechanism or mask acts as an intermediary between the contrast normalization process and anatomical structures. This intermediary layer allows the normalization algorithm to operate on the image while selectively shielding anatomical regions from excessive correction, thus maintaining both contrast consistency and structural accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Various methods and systems are provided for correcting contrast banding artifacts across multiple acquisitions in reconstructed images. In one embodiment, a method for computed tomography imaging comprises generating an original image comprising multiple subvolumes, segmenting the original image into different structures for each subvolume, selectively applying a mask-based correction through each area of subvolume that includes continuous structures to generate an updated image, and performing streak correction between the original image and the updated image to generate a final image. In this way, image quality may be improved without adjusting anatomical structures in an image.