Back to results
US 2009/0022380 A1Published application
Sub-voxel motion correction for phase-contrast magnetic resonance imaging
AssigneeVassol Inc.
InventorsMeide Zhao, Xiang He
Life of this patent
Estimating the term…
Abstract
A method and system are described for image registration using an information theoretic approach that can be used for correcting motion in blood flow studies as well as other applications. The joint probability distribution between two MRI (or other modality) images is estimated where the interpolation method is referred to as a continuous histogram.
Claims14
A method, comprising:
- obtaining intensity values for the N pixels (or voxels) making up each of a first image and a second image of a selected anatomical region, wherein each pixel intensity value is one of Q possible quantized intensity levels;
- geometrically transforming the pixel locations of the second image, referred to as the float image, in an attempt to put the second image in registration with the first image, referred to as the reference image;
- and,
- employing a joint probability density function as a part of an information theoretic measure of the degree to which the first and second images are in registration, wherein the joint probability density function is estimated by:
- computing an intensity function for the region around each pixel of each of the first and second images, referred to as a pixel region, wherein the intensity function is a continuous mapping of points within the pixel region to intensity values and wherein the intensity value of the pixel associated with the pixel region is a sample of the intensity function;
- assigning a bin to each possible pairwise combination of the Q quantized intensity values, referred to as an intensity value pair;
- for each pairwise combination of pixels from taken each of the first and second images, referred to as a pixel pair, updating each bin representing an intensity value pair that is mapped to by the corresponding intensity functions of the pixel pair;
- wherein, when a particular bin is updated by a particular pixel pair, a weighting factor reflective of the extent to which that pixel pair represents the intensity value pair for that bin is cumulatively added to the bin to thereby form a joint histogram;
- and,
- dividing the value of each bin in the joint histogram by the total number of pixel pairs, N2, to thereby form an estimate of the joint probability distribution function p(i,j) for the pixel intensities of the first and second images that gives the probability that randomly selected pixels from the first and second images have intensity values equal to i and j, respectively.
7 more elements
A system, comprising:
- an imaging scanner;
- an image processing computer configured to receive images generated by the imaging scanner, wherein the image processing computer is programmed to:
- obtain intensity values for the N pixels (or voxels) making up each of a first image and a second image of a selected anatomical region, wherein each pixel intensity value is one of Q possible quantized intensity levels;
- geometrically transform the pixel locations of the second image, referred to as the float image, in an attempt to put the second image in registration with the first image, referred to as the reference image;
- and,
- employ a joint probability density function as a part of an information theoretic measure of the degree to which the first and second images are in registration, wherein the joint probability density function is estimated by:
- compute an intensity function for the region around each pixel of each of the first and second images, referred to as a pixel region, wherein the intensity function is a continuous mapping of points within the pixel region to intensity values and wherein the intensity value of the pixel associated with the pixel region is a sample of the intensity function;
- assign a bin to each possible pairwise combination of the Q quantized intensity values, referred to as an intensity value pair;
- for each pairwise combination of pixels from taken each of the first and second images, referred to as a pixel pair, update each bin representing an intensity value pair that is mapped to by the corresponding intensity functions of the pixel pair, wherein, when a particular bin is updated by a particular pixel pair, a weighting factor reflective of the extent to which that pixel pair represents the intensity value pair for that bin is cumulatively added to the bin to thereby form a joint histogram;
- divide the value of each bin in the joint histogram by the total number of pixel pairs, N2, to thereby form an estimate of the joint probability distribution function p(i,j) for the pixel intensities of the first and second images that gives the probability that randomly selected pixels from the first and second images have intensity values equal to i and j, respectively;
- and,
- evaluate the registration of the images by an information theoretic measure using the computed joint probability distribution.
9 more elements
Dependencies are inferred from the claim text.