CT Image Reconstruction Using Displacement Fields Across Heart Cycles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
CT images reconstructed from inconsistent recording data due to patient movements, such as heartbeat or breathing, exhibit artifacts like blurred structures and discontinuities, which existing methods like high rotation speeds or dual source techniques fail to adequately address, especially in extended recording regions where data from different heart cycles are inconsistent.
Innovation Solution
A method involving the reconstruction of CT images using displacement vectors to align overlapping partial image volumes, interpolating a displacement vector field to correct for movement artifacts, ensuring seamless integration of data from multiple heart cycles, and generating an output image dataset without discontinuities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If data from multiple heart cycles are combined to improve coverage along the CT axis, then the coverage is improved, but discontinuities and artifacts arise at stack boundaries due to spatial differences between heart cycles
Solution Approach 1:
The patent segments the CT recording data into multiple stacks, each corresponding to a single heart cycle. This segmentation allows independent reconstruction of each stack from consistent temporal data, avoiding the discontinuities that arise when combining data from multiple heart cycles. The segmentation principle resolves the contradiction by prioritizing temporal consistency within each segment over continuous spatial coverage across segments.
Solution Approach 2:
The patent applies local quality by ensuring that each local stack region is reconstructed from data of a single heart cycle, guaranteeing high image quality and consistency within that local region. This local optimization approach accepts that coverage may be incomplete in some areas, but ensures that where coverage exists, the image quality is high and free from artifacts.
2Speed
If high rotation speeds are used to reduce the recording time span, then temporal resolution is improved, but artifacts remain due to small length scales of changes within the time interval
Solution Approach 1:
The patent utilizes the periodic nature of the cardiac cycle by synchronizing the recording and reconstruction process to specific phases of the heart cycle using ECG gating. By reconstructing images from data acquired during identical phases across multiple rotations, the method eliminates motion artifacts while maintaining the benefits of high rotation speed for temporal resolution.
Solution Approach 2:
The patent employs ECG signal feedback to control the reconstruction process. The ECG signal provides real-time information about the heart cycle phase, allowing the system to selectively combine data from different rotations only when they correspond to the same physiological phase, thereby eliminating artifacts while maintaining high rotation speeds.
3Area of stationary object
If table feed is used during recording to achieve complete coverage, then coverage is improved, but data from the same recording time point is not present for all positions along the CT axis
Solution Approach 1:
The patent segments the volumetric dataset into multiple axial stacks, where each stack is independently reconstructed from data of a single heart cycle. This segmentation approach allows the system to handle table feed motion by ensuring that each stack contains data from a consistent temporal reference point, thereby maintaining data consistency despite the continuous table movement during recording.
Solution Approach 2:
The patent performs preliminary sorting and organization of the raw CT data according to heart cycle phase and axial position before reconstruction. This preliminary action ensures that when data is later assembled into stacks, each stack contains only data from the appropriate temporal reference point, eliminating the inconsistency problem caused by table feed motion.
Data Source
AI summary
A method for reconstructing CT images, comprises: providing CT recording data; reconstructing overlapping partial images; establishing displacement vectors for registering overlap regions of the partial images; interpolating a displacement vector field for each partial image from associated sets of the displacement vectors of the two side regions; creating an output image dataset based on the CT recording data and the displacement vector fields; and outputting the output image dataset.