Improved spatial-spectral analysis by augmented modeling of 3D image appearance characteristics with application to radio frequency tagged cardiovascular magnetic resonance (CMR) (12030; 12061)
- 詳細技術說明
- None
- *Abstract
-
Tagged Cardiac Magnetic Resonance imaging (CMR) is a technique for detailed and non-invasive visualization of myocardium motion and deformation, with full 3D spatial geometric concordance. Local diseases, such as coronary atherosclerosis, as well as global diseases, such as heart failure and diabetes, result in wall dysfunction that manifests on tagged images. Magnetic Resonance (MR) tagging places a prespecified pattern of virtual temporary markers (called tags) inside soft body tissues such that tag lines are created by patterns of magnetic spin in the examined tissue, and motion in the tagged tissue can be measured from the images. This technique complements traditional anatomical images and can capture detailed information about the heart over time. The tag lines allow for the computation of displacement, velocity, rotation, elongation, strain, and twist of the heart. While traditional MR techniques carry only information about the motion at the boundaries of an object, the tag lines facilitate examination of the strain and displacement of the tissue interior in close detail.
The invention provides a computer aided processing system and automated method for processing tagged magnetic resonance (MR) images to generate improved tagged images in the spectral domain. Embodiments of the invention may process MR images to generate improved tagged images including reduced noise across one or more tag lines, augmented gradients across a tag profile, and/or amplified tag-to-background contrast. Embodiments of the invention may process Cardiac Magnetic Resonance (CMR) images modeling a distribution of CMR signals for the images corresponding to a cardiac cycle with an adaptive linear combination of discrete Gaussians (LCDG), where such modeled signals may be separated into signals corresponding to tag lines and signals corresponding to background. In addition, the images may be modeled by analyzing the images using a translation/rotation-invariant three-dimensional (3D) Markov-Gibbs random field model, where the images may be modeled by considering the CMR signals as voxel signals with multiple pairwise spatiotemporal (in-plane space and time) interactions. The gray-level signal of voxel-wise pairs may be analyzed to determine Gibbs potentials for interacting voxel pairs in the image set. A three dimensional energy function may be determined based at least in part on the interaction of the voxel pairs and the determined Gibbs potentials. The energy function may be utilized to amplify homogeneity of and the contrast between signals corresponding to tags and signals corresponding to background. Images processed using embodiments of the invention may provide more accurate spectral images for spectral analysis utilizing one or more spectral analysis methods.
Publication
Journal of Cardiovascular Magnetic Resonance 2012, 14(Suppl 1):P258 (1 February 2012)
Intellectual Property
U.S. Patent App. No. 13/834,304
U.S. Copyright Registration No. TX 7-546-226 for "MetaHARP Source Code"
CONTACT
For additional information, please contact University of Louisville's Office of Technology Transfer:
Telephone: 502-852-2965
Email: thinker@louisville.edu
Website: http://louisville.edu/thinker
ULRF Ref. Nos. 12030; 12061
- 國家/地區
- 美國
