Image analysis – Applications – Biomedical applications
Reexamination Certificate
1999-05-27
2002-05-28
Johns, Andrew W. (Department: 2621)
Image analysis
Applications
Biomedical applications
Reexamination Certificate
active
06396939
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to medical imaging. More specifically, the present invention is concerned with a segmentation method for medical images of structures such as, for example, bones.
BACKGROUND OF THE INVENTION
Most standard medical scanning devices, for example CT (Computerized Tomography) and MRI (Magnetic Resonance Imagery) scanners produce sets of two dimensional images that represent slices of the imaged anatomical structures. By successively examining these slices, a physician can build, in his mind, a three dimensional representation of the anatomical structure, from which a pathology can then be identified.
Recent advances in the field of medical imaging and computer graphics now allow the computation and the visualization of accurate three dimensional models of anatomical structures. In order to obtain these three dimensional models, every pixels that correspond to the desired structure must be identified in each two dimensional image, which is a long and tedious process given the large number of images that are generally required to obtain an accurate three dimensional representation of an anatomical structure.
The segmentation process, i.e. the identification of the desired pixels, is in general a complex process since some noise usually remains after applying standard image processing techniques, such as thresholding and noise reduction filtering. An additional challenge comes from the fact that some anatomical structures contain cavities or empty regions that must not be filled or merged with adjacent structures during the segmentation process.
According to a conventional segmentation process, a user reviews each image, one by one, using interactive image processing tools, such as thresholding, painting, and filling, in order to identify to which anatomical structure each pixel belongs. Accordingly, this process is long and tedious, and usually requires hours before the three dimensional model of the desired anatomical structure can be obtained.
A semi-automatic segmentation method is proposed in the international PCT application WO 98/39736, filed on Mar. 3, 1998 and naming HIBBARD as the inventor. This method includes the steps of:1) providing digital images corresponding to slices of a structure; 2) drawing by a user of a polygon within a ROI (Region Of Interest) or substructure to be segmented; 3) expanding the polygon by iteratively testing pixels of the images outside of, but adjacent to, the pixels that the polygon currently subtends. Pixels will be added to the polygon if the value of a decision rule function has a predetermined value. The perimeter of the polygon is considered the boundary of the substructure once it is found that none of the perimeter pixels satisfy the decision rule. The last step of the method is to compute the contour of the segmented substructure based on the boundary.
A drawback of Hibbard's method is that the user input involved in step 2 must be done for every slice (image) of the structure, therefore making the segmentation method time consuming. Moreover, to obtain an accurate contour, the user may have to perform step 2 a second time to provide new boundary segments for the polygon, depending of the complexity of the substructure.
More automated and reliable segmentation method and system are thus desirable.
OBJECTS OF THE INVENTION
An object of the present invention is therefore to provide a segmentation method and system free of the above described drawbacks.
Another object of the invention is to provide a reliable segmentation method and system that need less user interactions than the conventional methods and systems of the prior art.
SUMMARY OF THE INVENTION
More specifically, in accordance with the present invention, there is provided a segmentation method, comprising:
receiving a plurality of digital images representing consecutive slices of a structure that includes a substructure;
thresholding the plurality of images;
dividing the plurality of images in at least one group of images;
for each of at least one group of images,
a) estimating a substructure center position corresponding to the group;
b) using the substructure center position to find pixels of the substructure common to all images in the group;
c) for each image in the group,
c1) using the group substructure center position and the substructure common pixels to estimate a substructure center position corresponding to the image; and
c2) using the image substructure center position to remove noise both outside and inside the substructure, thereby producing segmented images.
In accordance with an aspect of the present invention, there is also providing a segmentation system comprising:
a first storing device for receiving a plurality of digital images representing consecutive parallel slices of a structure that includes a substructure;
a computer for thresholding the plurality of images;
dividing the plurality of images in at least one group of images;
for each of said at least one group of images,
a) estimating a substructure center position corresponding to the group;
b) using said substructure center position to find pixels of the substructure common to all images in the group;
c) for each image in the group,
c1) using said group substructure center position and said substructure common pixels to estimate a substructure center position corresponding to the image; and
c2) using the image substructure center position to remove noise both outside and inside the substructure, thereby producing segmented images.
In accordance with an aspect of the present invention, there is also providing an article of manufacture comprising:
a computer usable medium having a computer readable code means embodied in the medium for segmentation of images, the computer readable program code in the article of manufacture comprising:
computer readable program code means for causing the computer to receive a plurality of digital images representing consecutive parallel slices of a structure; the structure including a substructure;
computer readable program code means for causing the computer to threshold the plurality of images;
computer readable program code means for causing the computer to divide the plurality of images in at least one group of images;
computer readable program code means for causing the computer to estimate a substructure center position corresponding to each of at least one group images;
computer readable program code means for causing the computer to use the substructure center position to find pixels of the substructure common to all images in each group of at least one group of images;
computer readable program code means for causing the computer to use the group substructure center position and the substructure common pixels to estimate a substructure center position corresponding to each image in at least one group of images; and
computer readable program code means for causing the computer to use the image substructure center position to remove noise both outside and inside the substructure, thereby producing segmented images.
It is to be noted that in the following the terms pixel and point will both be used to described the smallest distinguishable element of a digital image.
Other objects, advantages and features of the present invention will become more apparent upon reading of the following non restrictive description of preferred embodiments thereof, given by way of example only with reference to the accompanying drawings.
REFERENCES:
patent: 4516261 (1985-05-01), Harding et al.
patent: 4630203 (1986-12-01), Szirtes
patent: 4835688 (1989-05-01), Kimura
patent: 4875165 (1989-10-01), Fencil et al.
patent: 4879652 (1989-11-01), Nowak
patent: 4882679 (1989-11-01), Tuy et al.
patent: 4914589 (1990-04-01), Crawford
patent: 4953087 (1990-08-01), Crawford
patent: 5056146 (1991-10-01), Nishide
patent: 5331553 (1994-07-01), Muehllehner et al.
patent: 5381518 (1995-01-01), Drebin et al.
patent: 5559712 (1996-09-01), Kihara et al.
patent: 5568384 (1996-10-01), Robb et al.
patent: 5782762 (1998-07-01), Vining
patent: 58253
Hu Tao
Labonté Francoys
Johns Andrew W.
Nakhjavan Shervin
Orthosoft Inc.
Wells St. John P.S.
LandOfFree
Method and system for segmentation of medical images does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for segmentation of medical images, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for segmentation of medical images will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2876460