Image analysis – Applications – Biomedical applications
Reexamination Certificate
2001-04-03
2004-11-02
Couso, Yon J. (Department: 2621)
Image analysis
Applications
Biomedical applications
C382S285000, C382S308000
Reexamination Certificate
active
06813373
ABSTRACT:
BACKGROUND OF THE INVENTION
The present application relates to diagnostic medical imaging. The invention finds particular application in segmenting and quantifying three-dimensional medical image data. It is to be appreciated however, that the present application finds further utility in quickly and accurately converting a series of two-dimensional images into a 3-D representation with accurately defined interfaces between distinct layer.
Quantification of the brain's gray matter, cortical thickness, the atrophy index of sulci in the cortex, and the principal curvatures of the cortex are important and clinically significant for analyzing brain diseases such as Alzheimer's, epilepsy, band heterotopia, dysplasia, schizophrenia, and dementia. The measurements can be computer estimated by accurately segmenting the cortex of the human brain. The above quantification tools also help in analyzing functional brain information such as in functional MRI (fMRI).
Typically, volumetric images are acquired and manipulated as a series of closely spaced, parallel two-dimensional image slices. Rendering this large database of sequential images into three dimensions is prohibitively expensive in computational power and in time. For example, the cerebral sheet at the interface between gray matter (GM) and white matter (WM) in the human brain is about 2500 cm
2
. The shape and convolutedness of the interface is important in many diagnostic procedures. For example, in neurological surgical planning
avigation and image guided surgery, an objective is to help the neuro-surgeon to plan the neuro-surgery in an operating room to identify the location of the tumor(s) in 3-D coordinates and the critical trajectories through which the tumors are accessed and removed. In surgical navigation, the objective is to navigate using the tools (such as the Y-probe) towards the target (such as a tumor in the brain) while avoiding critical brain structures and their functions during surgery. In Image Guided Surgery, the reconstructed patient's brain anatomy helps surgeons to interactively guide the surgery.
However, while the importance of cortical segmentation has been identified, it has not been easy to develop robust and high speed 3-D image creation and manipulation techniques.
The present invention contemplates an improved method and apparatus which overcomes the above-referenced problems and others.
SUMMARY OF THE INVENTION
In one embodiment of the present invention, a medical imaging method includes retrieving image data for a volume of interest, where the image data comprises a plurality of two-dimensional image slices. The plurality of image slices are transformed via gray scale mathematical morphology open-close and Bottom Hat edge preserving smoothing system into a three-dimensional volume. Interfaces within the three-dimensional volume are then identified.
In accordance with another aspect of the present invention, the transforming comprises applying a mathematical morphology transform to a selected one of the plurality of two-dimensional image slices.
In accordance with another aspect of the present invention, the mathematical morphology transform comprises performing an open smoothing operation and a closing smoothing operation on the selected two-dimensional image slice yielding an open slice and a closed slice. The open slice and the closed slice are then averaged using shifting.
In accordance with another aspect of the present invention, the mathematical morphology transform further comprises applying a Bottom-Hat smoothing system to the averaged slice.
In accordance with another aspect of the present invention, Bottom-Hat edge preservation and smoothing system comprises of a combination of histogramming the gray scale slices and applying a set of image algebra, logical Boolean operators and recursive mathematical morphology.
In accordance with another aspect of the present invention, recursive mathematical morphology comprises of thinning operation on binary images using skeletonization of binary shapes.
In another with another aspect of the present invention, the complete system comprises of linking the edge-preserved image denoising system with image derived constrained morphing system.
In accordance with another aspect of the present invention, the identifying comprises deriving constrained image forces which navigate the deformation process from the three-dimensional volume.
In accordance with another aspect of the present invention the identifying further comprises estimating fields within the three-dimensional volume contained by respective narrow bands.
In accordance with another aspect of the present invention the identifying further comprises under the derived image force constraints and within the narrow bands, propagating the estimated fields yielding updated fields, and checking the updated fields for convergence.
In accordance with another aspect of the present invention, the identifying further comprises estimating an isosurface under a morphological edge-preserved level set framework for each of the updated fields.
In accordance with another aspect of the present invention, the identifying further comprises rendering the updated fields.
In accordance with another aspect of the present invention, the estimating comprises generating a signed distance transform from initial fields or spheres.
In accordance with another aspect of the present invention, the generating comprises obtaining a vertex on one of the spheres, and computing polyline distances in the narrow band to yield accurate distances from the vertex.
In accordance with another aspect of the present invention, a medical image processing method comprises non-linear smoothing and edge-preserving a volume of interest represented by a series of spatially offset images, and segmenting embedded shapes in the smoothed volume of interest.
In accordance with another aspect of the present invention, the smoothing comprises sequentially selecting one of the series of spatially offset images. The selected image is smoothed with an open-close method and transformed with an ensemble of operators: morphology, logical operators, image algebra operators that constitute a bottom-hat morphological transform.
In accordance with another aspect of the present invention, the dynamic segmenting comprises estimating initial surfaces in the volume of interest, propagating the initial surfaces into revised surfaces, extracting an iso-surface from the revised surfaces, reinitializing the surfaces, checking for completion, and if incomplete, replacing previous surfaces with the reinitialized surfaces and repeating.
In accordance with another aspect of the present invention, the segmenting comprises deriving constrained morphing forces, computing an initial distribution model, morphing the distribution model by applying the derived forces, and extracting a volume from the morphed distribution model.
In accordance with another aspect of the present invention, the deriving comprises estimating a three-dimensional field distribution from initial spheres, computing data such as normals and offsets for selected voxels in the estimated three-dimensional field distribution, and determining a likelihood probability for a distribution such as WM or GM.
In accordance with another aspect of the present invention, the computing comprises computing a perpendicular distance between a point on an inner sphere and a corresponding point on an outer sphere.
In accordance with another embodiment of the present invention, a medical imaging apparatus comprises means for open smoothing and close smoothing a selected image slice to yield an open slice and a closed slice, means for averaging the open slice and the closed slice, and means for smoothing the averaged slice into a volume image representation.
In accordance with another aspect of the present invention, the imaging apparatus further comprises means for estimating spheres in the volume image representation, means for propagating the estimated spheres constrained by image force constraints and narrow ba
Liu Kecheng
Reden Laura M.
Suri Jasjit S.
Couso Yon J.
Koninklijke Philips Electronics , N.V.
LandOfFree
Image segmentation of embedded shapes using constrained... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image segmentation of embedded shapes using constrained..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image segmentation of embedded shapes using constrained... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3286824