6533b830fe1ef96bd1297285
RESEARCH PRODUCT
Image-based MRI Gradient Estimation
G. GarboRoberto AcquavivaStefano Mangionesubject
RegistrationImage qualityComputer scienceBiomedical EngineeringBiophysicsImage enhancement/restoration (noise and artifact reduction)Signal-To-Noise Ratio01 natural sciencesGeometric distortionImaging phantom030218 nuclear medicine & medical imaging03 medical and health sciences0302 clinical medicine0103 physical sciencesHomogeneity (physics)Gradient estimationRadiology Nuclear Medicine and imagingComputer visionMagnetic resonance imaging (MRI)010306 general physicsPhantoms Imagingbusiness.industrySettore ING-INF/03 - TelecomunicazioniImage EnhancementMagnetic Resonance ImagingMagnetic fieldMagnetArtificial intelligencebusinessAlgorithmsImage baseddescription
In order to reduce geometric distortion phenomena in MR images, every MRI system main magnet undergoes a shimming process. Since this process aims at optimizing magnetic field homogeneity within a so-called uniformity sphere, image quality outside this sphere is neglected. Since the fields vary smoothly in space, MR signal-to-noise ratio is still non-zero just outside the uniformity region, but correction of MR image distortion fails due to lack of magnetic field knowledge outside it. We propose a novel algorithm for measuring all the fields involved in the generation of images. Our proposal is based on exploitation of the distortion which can be observed in images of a known phantom. The proposed method will enable measurement of the fields in a region that can be bigger than the uniformity sphere depending on the phantom dimensions.
year | journal | country | edition | language |
---|---|---|---|---|
2017-12-19 |