Noise Correction in Diffusion MRI Using Iterative Fitting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Low signal-to-noise ratio (SNR) in diffusion-weighted magnetic resonance images poses challenges for accurate apparent diffusion coefficient (ADC) quantification, leading to errors in ADC maps, especially in high b-value images where noise can drop below the noise floor, affecting diagnostic accuracy.
Innovation Solution
A method that accounts for noise effects in MR images by using an iterative fitting algorithm to calculate initial and final values of selected variables, such as ADC, in a processor, which corrects noise in actual MR images, producing accurate quantitative measurements, and can be applied to various MR imaging techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If diffusion-weighted images are acquired with high b-values to improve diffusion contrast, then the ability to characterize tissue diffusion properties is improved, but the signal-to-noise ratio deteriorates causing noise to drop below the noise floor
Solution Approach 1:
The method performs preliminary noise characterization by acquiring images with multiple different noise levels (through repeated measurements with same parameters) before performing the diffusion quantification. This allows the system to establish the noise distribution characteristics in advance, which are then used to correct the ADC calculation and prevent noise floor effects from degrading measurement accuracy.
Solution Approach 2:
The method changes the approach from directly using noisy image intensities to calculating ADC values by fitting a signal model that explicitly accounts for noise characteristics. By transforming the problem from direct measurement to model-based estimation with noise parameters, the system can extract accurate diffusion information even when individual images have low signal-to-noise ratio.
2Reliability
If multiple measurements of diffusion-weighted images are averaged to improve signal-to-noise ratio, then the noise level is reduced, but the original noise distribution is altered affecting ADC accuracy
Solution Approach 1:
Instead of averaging images first and then calculating ADC (which alters noise distribution), the method inverts the approach by using individual measurements with their original noise distributions preserved, and applying noise correction during the ADC calculation process. This maintains the statistical properties of the noise while still achieving accurate diffusion quantification.
Solution Approach 2:
The method replaces the mechanical averaging operation with a statistical modeling approach. Rather than physically combining image data through averaging, the system uses a signal model that mathematically accounts for noise effects, substituting the mechanical process with a computational model that preserves the original data characteristics while achieving noise robustness.
3Productivity
If conventional fitting algorithms are used to calculate ADC values from noisy images, then the calculation process is simple and fast, but the ADC maps contain errors due to uncorrected noise effects
Solution Approach 1:
The method performs preliminary characterization of noise effects by analyzing multiple measurements with identical parameters before performing the final ADC calculation. This preliminary analysis establishes the noise distribution and correction factors that are then applied during the fitting process, allowing accurate ADC values to be obtained without requiring complex iterative corrections during the main calculation.
Solution Approach 2:
The method introduces noise parameters as intermediary variables in the fitting process. Rather than directly fitting ADC values to noisy image intensities, the system introduces noise characteristics as intermediate parameters that mediate between the raw data and the final ADC calculation, allowing the fitting algorithm to account for noise effects while maintaining computational efficiency.
Data Source
AI summary
In a method to correct noise effects in magnetic resonance (MR) images, which is executed in a processor (computer), the processor executes a fitting algorithm in order to calculate initial values for each of selected variables in signal model that models noise effects in a modeled, noise-containing MR image. The processor then iteratively executes the same or a different fitting algorithm, in order to generate final values for each of the selected variables. The processor is provided with an actual, acquired MR image that contains noise, and the processor uses the final values of the selected variables to calculate synthetic signal intensities in the MR image, thereby producing a synthetic MR image with no noise bias effects of errors. This synthetic image is made available in electronic form at an output of the processor, as a data file.