# US 10,768,260 B2 · System and method for controlling noise in magnetic resonance imaging using a local low rank technique > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,768,260 B2 - Document ID: US10768260B2 - Status: Granted - Kind: B2 - Application number: 16/133,550 - Filed: 2018-09-17 - Published: 2019-03-21 - Granted: 2020-09-08 - Earliest claimed: 2017-09-18 - Assignees: Regents of the University of Minnesota - Applicants: Regents of the University of Minnesota - Inventors: Steen Moeller; Mehmet Akcakaya - Web page: https://patentgrep.com/patent/US10768260B2 ## CPC - G01R 33/565: Correction of image distortions, e.g. due to magnetic field inhomogeneities - G06T 5/50: using two or more images, e.g. averaging or subtraction - G06T 5/70: Denoising; Smoothing - G06T 7/0012: {Biomedical image inspection} - G06T 12/10: Image preprocessing, e.g. calibration, positioning of sources or scatter correction - G01R 33/5611: {Parallel magnetic resonance imaging, e.g. sensitivity encoding [SENSE], simultaneous acquisition of spatial harmonics [SMASH], unaliasing by Fourier encoding of the overlaps using the temporal dimension [UNFOLD], k-t-broad-use linear acquisition speed-up technique [k-t-BLAST], k-t-SENSE (structural details of arrays of sub-coils G01R33/3415)} - G06T 2207/10016: Video; Image sequence - G06T 2207/10088: Magnetic resonance imaging [MRI] - G06T 2207/20182: Noise reduction or smoothing in the temporal domain; Spatio-temporal filtering - G06T 2207/30024: Cell structures in vitro; Tissue sections in vitro - G01R: MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES (indicating correct tuning of resonant circuits H03J3/12) - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract A system and method for controlling noise in magnetic resonance imaging (MRI) are provided. In one aspect, the method includes reconstructing a series of images of the target using the image data, with each image being defined using signal-to-noise (SNR) units, and selecting an image patch corresponding to the series of images. The method also includes forming a matrix by combining vectors generated using the image patch, and applying a local low rank denoising technique using the matrix and the series of images to generate at least one denoised image. ## Claims ### Claim 1 A method for generating a denoised image from data acquired with a magnetic resonance imaging (MRI) system, the method including steps comprising: (a) controlling an MM system to acquire magnetic resonance (MR) image data from a target; (b) reconstructing a series of MR images of the target from the MR image data using a computer system, wherein pixel values in each MR image in the series of MR images are defined using signal-to-noise ratio (SNR) units; (c) selecting with the computer system, an image patch corresponding to the series of MR images; (d) forming a matrix with the computer system by combining vectors generated using the image patch; and (e) applying a local low-rank denoising technique with the computer system using the matrix and the series of MR images to generate at least one denoised MR image. ### Claim 2 (depends on 1) The method of claim 1, wherein the MR image data acquired from the target comprises at least one of functional MM (fMRI) image data, diffusion image data, arterial spin labeling (ASL) image data, or parameter imaging data. ### Claim 3 (depends on 1) The method of claim 1, wherein reconstructing the series of MR images in the step (b) comprises utilizing a noise-decorrelation technique. ### Claim 4 (depends on 1) The method of claim 1, wherein reconstructing the series of MR images in the step (b) comprises using a parallel image reconstruction. ### Claim 5 (depends on 4) The method of claim 4, wherein the parallel image reconstruction is a k-space based parallel image reconstruction. ### Claim 6 (depends on 5) The method of claim 5, wherein the k-space based parallel image reconstruction is a Generalized Autocalibrating Partially Parallel Acquisition (GRAPPA) reconstruction. ### Claim 7 (depends on 4) The method of claim 4, wherein the parallel image reconstruction is an image domain-based parallel image reconstruction. ### Claim 8 (depends on 7) The method of claim 7, wherein the image domain-based parallel image reconstruction is a sensitivity-encoding (SENSE) reconstruction. ### Claim 9 (depends on 4) The method of claim 4, further comprising generating a g-factor map based on the parallel image reconstruction and normalizing the series of MR images using the g-factor map. ### Claim 10 (depends on 9) The method of claim 9, further comprising normalizing the at least one denoised MR image using the g-factor map in order to convert the at least one denoised image to the SNR units. ### Claim 11 (depends on 4) The method of claim 4, further comprising generating a g-factor map based on the parallel image reconstruction and normalizing the series of MR images using the g-factor map in order to convert the series of MR images to the SNR units. ### Claim 12 (depends on 1) The method of claim 1, wherein the local low-rank denoising technique implements a singular value decomposition. ### Claim 13 (depends on 1) The method of claim 1, wherein the local low-rank denoising technique implements a singular value thresholding. ### Claim 14 (depends on 13) The method of claim 13, wherein the singular value thresholding is implemented using a threshold value computed based on singular values of a random Gaussian matrix. ### Claim 15 (depends on 1) The method of claim 1, wherein the step (c) comprises extracting blocks of pixels from the series of MR images that are centered on a pixel location. ### Claim 16 (depends on 1) The method of claim 1, wherein the step (c) is repeated to select a plurality of image patches having different patch sizes and the local low-rank denoising technique is a multiscale denoising implemented on the plurality of image patches. ### Claim 17 (depends on 1) The method of claim 1, wherein using the matrix in the step (e) comprises using a distance weighted approach to combine overlapping image blocks. ### Claim 18 (depends on 1) The method of claim 1, wherein the matrix is a Casorati matrix. ### Claim 19 (depends on 1) The method of claim 1, wherein the steps (c) through (e) are repeated using image patches at different scales and a stopping criteria.