DSA Image Correction Using Two-Stage Registration and Buffer Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital subtraction angiography (DSA) technologies suffer from low image accuracy in regions of interest due to independent motion modes, leading to motion artifacts that cannot be accurately corrected through conventional registration methods.
Innovation Solution
A method involving two-stage registration and buffer area processing is employed, where a first registration is performed on a scan image and a reference image, followed by a second registration on the region of interest, with a buffer area calculation and transition processing to eliminate artifacts, resulting in a corrected scan image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single registration is performed on the entire image, then the processing time is reduced and the overall image structure is maintained, but the local region of interest cannot be accurately corrected due to independent motion modes
Solution Approach 1:
The patent divides the registration process into two stages: first registering the entire image to obtain a global transformation, then registering only the region of interest using the global result as initialization. This segmentation allows accurate local correction while reducing overall complexity compared to performing separate registrations for each region.
Solution Approach 2:
The patent applies different registration strategies to different regions: global registration for the entire image to establish baseline alignment, and local registration specifically for the region of interest to correct independent motion. This local quality approach ensures accurate correction where needed while maintaining efficiency elsewhere.
2Measurement precision
If registration is performed on the entire image, then the overall alignment is achieved, but motion artifacts remain in the region of interest due to independent motion modes
Solution Approach 1:
The patent segments the registration into global and local stages. Global registration handles overall alignment, while local registration specifically addresses motion artifacts in the region of interest. This segmentation enables removal of motion artifacts without sacrificing overall alignment accuracy.
Solution Approach 2:
The patent performs global registration as a preliminary action to establish a baseline alignment and transformation matrix. This preliminary registration then serves as the foundation for the subsequent local registration, allowing the system to build upon the global alignment to specifically correct local motion artifacts.
3Object-generated harmful factors
If a buffer area is introduced between the region of interest and surrounding areas, then transition processing eliminates artifacts at boundaries, but the processing time and computational complexity increase
Solution Approach 1:
The patent applies buffer area processing locally only at the boundaries between the region of interest and surrounding areas, rather than processing the entire image. This localized approach to artifact elimination reduces unnecessary computational time while effectively addressing boundary artifacts where they occur.
Solution Approach 2:
The patent introduces a buffer area as an intermediary zone between the region of interest and surrounding areas. This buffer serves as a transition region that facilitates smooth blending and eliminates abrupt artifacts at boundaries. The buffer area acts as a mediator that reconciles the different registration characteristics of the region of interest and surrounding areas.
Data Source
AI summary
A method for correcting a digital subtraction angiography image is disclosed. The method includes: performing a first registration on a first scan image and a reference image to obtain a second scan image; performing a second registration on a region of interest of the second scan image and a region of interest of the reference image to obtain a third scan image; calculating a buffer area between the third scan image and the second scan image, and performing transition processing on the buffer area, to obtain a processed buffer area image; and fusing the second scan image, the third scan image, and the processed buffer area image to obtain a corrected scan image. A computer device and a non-volatile computer readable storage medium are also disclosed.