MRI Denoise Strength Based on G-Factor Map
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical image processing techniques for MRI fail to effectively improve diagnosis accuracy by not adequately addressing noise removal and image mixing based on RF coil sensitivity and g-factor distributions, leading to potential loss of diagnostic structures and uneven noise distribution.
Innovation Solution
A medical image processing apparatus that determines denoise strength and mixing rates based on coil sensitivity and g-factor maps, generating a mixed image by combining original and denoised images using a blend rate map, thereby enhancing diagnosis accuracy and reducing noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If denoise processing is applied to MRI data, then noise is reduced, but diagnostic structures may be lost and diagnosis accuracy deteriorates
Solution Approach 1:
The patent applies different denoise strengths to different regions of the image based on local g-factor values. Regions with high g-factor (where noise is more prominent) receive stronger denoise processing, while regions with low g-factor (where diagnostic structures are preserved) receive weaker processing. This local adaptation resolves the contradiction by making denoise intensity spatially variable rather than uniform.
Solution Approach 2:
The patent dynamically adjusts the denoise strength parameter based on the g-factor map. By changing the denoise parameter according to local g-factor values, the system optimizes the balance between noise removal and structure preservation in different regions, thereby maintaining diagnosis accuracy while reducing noise.
2Device complexity
If uniform denoise strength is applied, then processing is simple, but noise distribution remains uneven and diagnosis accuracy is compromised
Solution Approach 1:
The patent introduces spatially varying denoise strength based on local g-factor characteristics. This local quality approach transforms uniform processing into adaptive processing, where each region receives appropriate denoise treatment, improving diagnosis accuracy without excessive complexity increase.
Solution Approach 2:
The patent performs preliminary calculation of the g-factor map before applying denoise processing. This preliminary action provides the basis for determining optimal denoise strength in each region, enabling adaptive processing that improves diagnosis accuracy while keeping the overall process manageable.
3Object-affected harmful factors
If strong denoise is applied, then noise is reduced, but image definition deteriorates and re-scanning may be needed
Solution Approach 1:
The patent applies strong denoise only in regions where g-factor indicates high noise content, while preserving image definition in regions with low g-factor. This localized approach prevents over-denosing that would blur diagnostic structures, thereby maintaining image definition while effectively reducing noise.
Solution Approach 2:
The patent replaces uniform mechanical denoise application with an adaptive system that uses g-factor information to guide denoise strength. This substitution of the denoise mechanism with an intelligence-driven approach prevents loss of image definition while achieving noise reduction.
Data Source
AI summary
A medical image processing apparatus of embodiments includes processing circuitry. The processing circuitry is configured to acquire a data and a g-factor map. The data is generated as a result of a reception by an RF coil. The processing circuitry is configured to determine strength of denoise performed on the data on the basis of the g-factor map and perform the denoise on the data.