Cardio CT Image Reconstruction Using Multi-Cycle Data Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CT reconstruction methods, such as Filtered Back Projection, face challenges with movement imprecision in cyclically moving objects, particularly evident in cardio CT examinations where heart movement introduces significant artifacts, leading to reduced image quality and temporal resolution.
Innovation Solution
A method that combines measurement data from multiple movement cycles of a cyclically moving object to reconstruct CT pictures, using difference information to compute a result picture that leverages the advantages of both high temporal resolution and noise reduction, achieved by forming weighted sums or averages of pixel values from different cycle data, and applying thresholding and filtering to the difference picture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If measurement data from multiple movement cycles is combined to reconstruct CT pictures, then temporal resolution is improved, but noise increases due to combining data from different cycles
Solution Approach 1:
The patent segments the image reconstruction process by dividing it into a first picture from single-cycle data and a second picture from multi-cycle data. This segmentation allows each picture to serve different purposes: the first maintains low noise while the second provides high temporal resolution, resolving the contradiction between these two competing requirements.
Solution Approach 2:
The patent applies different quality characteristics to different parts of the final result by comparing the first and second pictures. Regions with minimal movement artifacts from the single-cycle picture are preserved, while regions benefiting from the multi-cycle picture's temporal resolution are enhanced, creating a composite image with locally optimized quality.
2Reliability
If data from multiple movement cycles is combined for reconstruction, then movement artifacts are reduced, but calculation complexity increases
Solution Approach 1:
The calculation process is segmented into separate reconstruction steps: first reconstructing a picture from single-cycle data, then reconstructing another picture from multi-cycle data, and finally combining them through comparison. This segmentation simplifies the overall calculation complexity by breaking down the complex multi-cycle reconstruction into manageable stages.
Solution Approach 2:
The patent applies multi-cycle data combination selectively rather than universally. By using difference information to identify regions where movement artifacts are problematic, the method applies the computationally intensive multi-cycle reconstruction only where needed, rather than to the entire image, thus reducing overall calculation complexity.
3Shape
If high temporal resolution is achieved through single-cycle reconstruction, then image sharpness is maintained, but movement imprecision increases
Solution Approach 1:
The patent merges the advantages of single-cycle and multi-cycle reconstructions by combining the sharp, artifact-free regions from the single-cycle picture with the movement-corrected regions from the multi-cycle picture. This merging creates a final composite image that achieves both sharpness and movement precision simultaneously.
Solution Approach 2:
The comparison of the first and second pictures serves as an intermediary mechanism that identifies movement artifacts. This intermediary step allows the system to selectively replace artifact-containing regions with corrected data from the multi-cycle reconstruction, preserving sharpness while eliminating movement imprecision.
Data Source
AI summary
A method for reconstructing picture data of a cyclically-moving object from measurement data is disclosed, with the measurement data being detected beforehand for a relative rotational movement between a radiation source of a computed tomography system and the object under examination during a plurality of movement cycles of the object under examination. In at least one embodiment, a first picture and a second picture are determined from the measurement data, with measurement data of different movement cycles being combined for reconstruction of the second picture into a measurement dataset to be used as the basis for the picture reconstruction. Difference information is computed by comparing the first picture with the second picture. Using the difference information, a result picture is computed from the first picture and the second picture.