Automated Roadmap Weighting for Medical Image Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current roadmap methods in medical imaging face challenges with image registration due to patient and imaging apparatus movement, leading to motion artifacts and unsatisfactory performance, especially when large regions or different movements occur within sub-regions, requiring manual adjustments and compromising image evaluation.
Innovation Solution
An automated method for determining weighting images of anatomical and object images using prespecified weighting values and functions to prioritize relevant image regions, enabling accurate registration and overlaying by generating an overall weighting image that prioritizes pixels or regions of interest, thereby improving image quality and reducing manual intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pixel-shift algorithms are used to register images, then image registration is achieved, but performance becomes unsatisfactory when large regions are mapped or different movements occur in different sub-regions
Solution Approach 1:
The patent applies local quality by assigning different weighting values to different regions of the image. The weighting image has higher weights for regions of interest (where the medical object is located) and lower weights for other regions. This allows the pixel-shift algorithm to focus on accurately registering critical areas while being less stringent about other regions, thereby maintaining reliable performance even when large regions are mapped or different movements occur in different sub-regions.
2Ease of operation
If manual adjustment of settings is required during intervention, then workflow is adversely affected, but automatic methods lack flexibility for complex movements
Solution Approach 1:
The patent implements self-service by automatically generating the weighting image based on the anatomical image and object image. The system automatically identifies regions of interest and assigns appropriate weighting values without requiring manual intervention. This maintains workflow efficiency while simultaneously providing adaptability for complex movements, as the weighting image dynamically adapts to the specific intervention scenario.
3Device complexity
If uniform weighting is applied to all image regions, then processing is simplified, but registration accuracy suffers in critical areas during complex movements
Solution Approach 1:
The patent applies local quality by creating a weighting image with spatially varying weights. Regions containing the medical object or anatomical features of interest receive higher weighting values, while other regions receive lower weights. This approach maintains manageable processing complexity through automated generation while significantly improving registration accuracy in critical areas during complex movements.
Data Source
AI summary
A weighting for a roadmap method is automatically determined. A first or a second weighting image is generated from an anatomical image and an object image. For this purpose, a prespecified first weighting value is assigned to pixels belonging to a prespecified anatomical feature or to an instrument. Other pixels are assigned increasingly small weighting values at increasing distances from the anatomical feature or from the instrument toward an edge of a respective recording region according to a prespecified monotonously decreasing function in dependence upon the location. An overall weighting image is generated by combining the first and the second weighting images with one another and/or a region of interest determined using the overall weighting image are then provided as input data for an image processing algorithm.