MRI Phase Coding Artifact Elimination via Dual-Axis Image Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance images often suffer from phase coding artifacts due to patient movement, natural physiological activities, and signal sampling issues, which can lead to incorrect diagnoses by obscuring or distorting image details.
Innovation Solution
A method that involves obtaining two images with phase coding along columns and rows, respectively, and applying a processing sequence to compare and replace portions with different brightness intensity values, creating derived images that are further processed to eliminate artifacts by adjusting threshold values and averaging intensity values, ultimately producing a high-definition image without the need for operator-set phase coding directions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If operator adjusts operative parameters (filters, contrast, brightness) to attenuate artifacts, then artifact effects are reduced, but image may be altered and disease indicators may be hidden
Solution Approach 1:
The patent replaces manual operator adjustments (mechanical/subjective process) with an automated image processing system that uses algorithmic analysis and comparison of multiple images to eliminate artifacts while preserving diagnostic information
Solution Approach 2:
The system performs self-correction by automatically analyzing the images, identifying artifacts through comparison, and eliminating them without requiring operator intervention or subjective judgment
2Object-affected harmful factors
If operator applies filters or adjusts contrast and brightness to reduce artifacts, then artifact visibility is reduced, but the process requires consolidated experience and operator sensitivity
Solution Approach 1:
The system automates the artifact reduction process by having it self-analyze and self-correct through algorithmic comparison of multiple images, eliminating the need for operator experience or subjective judgment in adjusting parameters
3Object-affected harmful factors
If phase coding acquisition is performed along rows or columns to control artifact propagation direction, then artifact propagation is directed along specific directions, but the parameter must be manually set by operator before acquisition
Solution Approach 1:
The system automatically determines the optimal phase coding direction by analyzing the images and identifying artifact patterns, then proceeds with processing without requiring the operator to manually set the phase coding direction parameter before acquisition
4Object-affected harmful factors
If saturation bands are applied to regions outside anatomical part to prevent artifact generation, then artifact production from those regions is prevented, but the shape cannot be modified to adapt to irregular regions and time-consuming operator arrangement is required
Solution Approach 1:
Instead of preventing artifact generation through saturation bands, the patent extracts the problematic regions through comparative analysis of multiple images, identifying and eliminating artifacts post-acquisition without requiring pre-acquisition saturation band placement
Solution Approach 2:
The system automatically identifies artifact regions through image comparison and processing without requiring operator arrangement of saturation bands, making the process self-acting and eliminating manual configuration steps
Data Source
Figure 1
Figure 2~5
Figure 6
AI summary
A method for eliminating artifacts from a magnetic resonance image begins with the steps of obtaining a first image (100a), using a phase coding along columns, and a second image (100b), using a phase coding along rows. Both images (100a, 100b) comprise a plurality of portions (101), each having a determined intensity value i arranged in a matrix structure (m-n), i.e. comprising m rows and n columns. The images (100a, 100b) are subject to a processing sequence that provides a comparative analysis by columns carried out comparing the corresponding columns and calculating the number of different portions, i.e. having a different intensity average value (imi?im2). The images (100a, 100b), furthermore, are subject to an analysis by rows that provides a comparison between the corresponding rows and calculating the number of different portions, i.e. portions having a different intensity average values (imi?im2). This way, a first and a second derived image of I generation are obtained, improved with respect to the starting images [Fig. 1].