CT Image Registration Using MR Data for Respiration Phase Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The registration of CT and ultrasound images is hindered by differences in respiration phases, leading to registration errors and decreased image quality due to organ deformation, which existing methods like manual feature extraction and simulation-based registration are inefficient and inconvenient.
Innovation Solution
An image processing apparatus and method that uses MR image data to align CT images across different respiration phases by determining similarity based on anatomical features and applying displacement vectors to register CT and ultrasound images accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If CT and ultrasound images are registered without considering respiration phase differences, then the registration process is simple, but registration errors increase and image quality decreases due to organ deformation
Solution Approach 1:
The patent applies preliminary action by selecting MR image data that corresponds to the same respiration phase as the ultrasound image before performing registration. This pre-selection ensures that organ deformation due to respiration is minimized, thereby improving registration accuracy without requiring complex real-time respiration synchronization during the registration process itself.
Solution Approach 2:
The patent uses MR image data as an intermediary to bridge the gap between CT and ultrasound images. By selecting MR images that match the respiration phase of the ultrasound image and registering CT to these intermediate MR images, the system achieves more accurate multi-modal registration while avoiding direct complex registration between images taken at different respiration phases.
2Productivity
If manual feature extraction and simulation-based registration are used, then registration can be performed, but the process becomes inefficient and inconvenient
Solution Approach 1:
The patent applies self-service by using anatomical features from MR images to automatically determine similarity and guide the selection of appropriate MR image data for registration. This automated approach eliminates the need for manual feature extraction and simulation-based methods, thereby improving both registration efficiency and operational convenience while maintaining high accuracy.
3Quantity of substance
If MR image data from different patients is used, then more data is available for selection, but anatomical similarity decreases leading to poorer registration results
Solution Approach 1:
The patent applies local quality by evaluating and selecting MR image data based on anatomical similarity metrics specific to the patient's anatomy. Rather than using all available MR data uniformly, the system identifies and selects only those MR images that exhibit local anatomical characteristics matching the target patient, thereby ensuring high reliability while utilizing multiple data sources.
Data Source
AI summary
An image processing apparatus includes an input device configured to receive a computerized tomography image of a first respiration phase of an object; a data selector configured to select magnetic resonance image data similar to the object from among magnetic resonance image data; and an image generator configured to generate a computerized tomography image of a second respiration phase of the object based on the selected magnetic resonance image data.