# US 10,871,535 B2 · Magnetic resonance fingerprinting optimization in magnetic resonance imaging > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,871,535 B2 - Document ID: US10871535B2 - Status: Granted - Kind: B2 - Application number: 16/143,865 - Filed: 2018-09-27 - Published: 2019-03-28 - Granted: 2020-12-22 - Earliest claimed: 2017-09-28 - Assignees: Siemens Healthcare GmbH - Applicants: Siemens Healthcare GmbH - Inventors: Christian Alexander Braun; Mariappan S. Nadar; Xiao Chen; Boris Mailhe - Web page: https://patentgrep.com/patent/US10871535B2 ## CPC - G01R 33/56: Image enhancement or correction, e.g. subtraction or averaging techniques {, e.g. improvement of signal-to-noise ratio and resolution} - G01R 33/50: based on the determination of relaxation times {, e.g. T1 measurement by IR sequences; T2 measurement by multiple-echo sequences} - G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06N 3/092: Reinforcement learning - G06V 40/1365: {Matching; Classification} - G01R 33/4818: {MR characterised by data acquisition along a specific k-space trajectory or by the temporal order of k-space coverage, e.g. centric or segmented coverage of k-space} - G06N 3/126: Evolutionary algorithms, e.g. genetic algorithms or genetic programming - G01R: MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES (indicating correct tuning of resonant circuits H03J3/12) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract For determining a configuration for a MR scanner using MRF, MRF is optimized using a model of acquisition and reconstruction together. The effects of under sampling and reconstruction are included in the model being optimized with the sequence. A spatial distribution of response is used in the MRF optimization. An added transformation may be included to convert the fingerprint into a domain for noise removal and/or more direct parameter estimation without the dictionary. Artificial intelligence may be used to alter and/or select the type of optimization or optimization settings and/or to optimize. ## Claims ### Claim 1 A method for determining a configuration for a magnetic resonance scanner using magnetic resonance fingerprinting, the method comprising: generating a first image fingerprint from magnetic resonance corresponding to a first sequence and a plurality of ground truth values of magnetic resonance output parameters; transforming the first image fingerprint to k-space; applying a first sampling pattern to the transformed first image fingerprint; reconstructing first values of the magnetic resonance output parameters from results of the applying of the first sampling pattern, the reconstructing using a first reconstruction setting; scoring the first values with respect to the ground truth values; optimizing the scoring based on repetition of the generating with different sequences than the first sequence, transforming, applying with different sampling patterns than the first sampling pattern, and reconstructing with different reconstructing settings; selecting a magnetic resonance fingerprinting sequence from the first and different sequences, a magnetic resonance fingerprinting sampling pattern from the first and different sampling patterns, and a magnetic resonance fingerprinting reconstruction setting from the first and different reconstruction settings; and providing the selected magnetic resonance fingerprinting sequence, sampling pattern, and reconstruction setting for the magnetic resonance scanner. ### Claim 2 (depends on 1) The method of claim 1 wherein generating, transforming, applying, reconstructing, and scoring are performed for each location in a multi-dimensional distribution of locations such that the selecting is based on the scoring being for the multi-dimensional distribution of the locations. ### Claim 3 (depends on 1) The method of claim 1 wherein generating comprises generating with Block simulation. ### Claim 4 (depends on 1) The method of claim 1 wherein generating comprises generating with a phantom, the ground truth values being from a scan of the phantom. ### Claim 5 (depends on 1) The method of claim 1 wherein the repetition optimizes for the magnetic resonance fingerprinting sequence, the magnetic resonance sampling pattern, and the magnetic resonance reconstruction setting in sequence. ### Claim 6 (depends on 1) The method of claim 1 wherein scoring comprises scoring with a penalty for one of the first and different sequences, the first and different sampling patterns, or the first and different reconstruction settings. ### Claim 7 (depends on 1) The method of claim 1 wherein the repetition is limited by one or more constraints. ### Claim 8 (depends on 1) The method of claim 1 wherein the first and different sampling patterns are different trajectories, different amounts of undersampling, and/or different sampling orders. ### Claim 9 (depends on 1) The method of claim 1 wherein the first and different reconstruction settings comprise different types of reconstruction and/or different values of reconstruction parameters. ### Claim 10 (depends on 1) The method of claim 1 wherein reconstructing comprises inverse transforming the results and reconstructing by dictionary matching from the inverse transforming. ### Claim 11 (depends on 1) The method of claim 1 wherein scoring comprises scoring as a function of variance by location. ### Claim 12 (depends on 1) The method of claim 1 further comprising inverse transforming the results, applying a transform operator, and removing information in a transform operator domain, and wherein reconstructing comprises determining the first values from an output of removing the information in the transform operator domain. ### Claim 13 (depends on 1) The method of claim 1 wherein optimizing comprises optimizing with a particle swarm optimization. ### Claim 14 (depends on 1) The method of claim 1 wherein optimizing comprises optimizing with an optimization algorithm and/or optimization setting selected by a machine-learned network. ### Claim 15 (depends on 1) The method of claim 1 wherein optimizing comprises optimizing with a machine-learned network. ### Claim 16 A method for determining a configuration for a magnetic resonance scanner using magnetic resonance fingerprinting, the method comprising: optimizing the magnetic resonance fingerprinting with a model including (a) a scan sequence and/or a sampling pattern and (b) a reconstruction from fingerprint matching to a dictionary, a score for the optimizing being based on a reconstruction map of a plurality of locations, the optimizing being of (a) the scan sequence and/or a sampling pattern and (b) the reconstruction; providing a configuration as (a) the scan sequence and/or the sampling pattern and (b) the reconstruction as optimized to the magnetic resonance scanner. ### Claim 17 (depends on 16) The method of claim 16 wherein optimizing comprises optimizing the scan sequence and the reconstruction, and wherein providing comprises providing the configuration as the scan sequence as optimized and the reconstruction as optimized. ### Claim 18 (depends on 16) The method of claim 16 wherein optimizing comprises optimizing the sampling pattern, and wherein providing the configuration comprises providing the sampling pattern as optimized. ### Claim 19 (depends on 16) The method of claim 16 wherein optimizing comprises optimizing based on an output of a machine-learned network for optimization. ### Claim 20 (depends on 16) The method of claim 16 wherein the model applies a domain transformation with fingerprint matching being in a domain of the domain transformation.