System and Method of Automatic Prioritization and Analysis of Medical Images

• US 20090028403A1
• Filed: 02/18/2007
• Published: 01/29/2009
• Est. Priority Date: 03/03/2006
• Status: Active Grant

• Current Assignee: Medic Vision - Brain Technologies Ltd.
• Original Assignee: Medic Vision - Brain Technologies Ltd.
• Inventors: Dahan, Ellran, Devir, Zvi, Bar-Aviv, Ezer
• Granted Patent: US 8,331,637 B2
• Time in Patent Office: Days
• US Class Current: 382/128
• CPC Class Codes: G06T 2207/20128 Atlas-based segmentation
• CPC Class Codes: G06T 2207/20161 Level set
• CPC Class Codes: G06T 2207/30016 Brain
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/149 involving deformable models, e.g. active contour models
• CPC Class Codes: G06T 7/75 involving models
• CPC Class Codes: G06V 10/24 Aligning, centring, orientation detection or correction of the image
• CPC Class Codes: G06V 10/40 Extraction of image or video features
• CPC Class Codes: G06V 10/754 involving a deformation of the sample pattern or of the reference pattern; Elastic matching
• CPC Class Codes: G06V 2201/032 of protuberances, polyps nodules, etc.
• CPC Class Codes: G16H 15/00 ICT specially adapted for medical reports, e.g. generation or transmission thereof
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 70/60 relating to pathologies

Abstract

A system for analyzing a source medical image of a body organ. The system comprises an input unit for obtaining the source medical image having three dimensions or more, a feature extraction unit that is designed for obtaining a number of features of the body organ from the source medical image, and a classification unit that is designed for estimating a priority level according to the features.

Claims

1. A system for analyzing a source medical image of a body organ, said system comprising:

an input unit for obtaining said source medical image, said source medical image having at least three dimensions;

a segmentation unit, operatively associated with said input unit, configured for segmenting said source medical image by matching with a reference medical image of a respective normal organ, said reference medical image already being segmented, said reference medical image having at least three dimensions;

a feature extraction unit, operatively associated with said segmentation unit, configured for obtaining a plurality of features of said body organ from said segmented source medical image; and

a pathology detector unit, operatively associated with said feature extraction unit, configured for indicating at least one pathology based on said plurality of features.

4. The system of claim 1, wherein said body organ is brain.

5. The system of claim 1, further comprising a collection of normative values for said features.

6. The system of claim 5, wherein each segmented part of the image has at least one normative value associated with it.

7. The system of claim 6, wherein said features include at least four different features.

8. The system of claim 1, wherein at least some of said features relate to a symmetry of said organ and at least one of said features relates to a symmetry other than geometrical symmetry.

9. The system of claim 1, further comprising a registration module being configured for performing a spatial image registration according to spatial positioning differences between said body organ and said reference normal organ, said spatial positioning differences being used during by said segmentation unit for said segmenting.

10. The system of claim 9, wherein said registration module comprises a global motion sub-module being configured for estimating a global motion vector according to said spatial positioning differences, said global motion vector being used during said spatial image registration.

12. The system of claim 9, wherein said registration module comprises a 3D affine transformation sub-module being configured for estimating an affine transformation vector according to said spatial positioning differences, said affine transformation vector being used during said spatial image registration and wherein said reference medical image is associated with a plurality of reference descriptors, said 3D affine transformation sub-module being configured for identifying a plurality of source descriptors in said source medical image said, said spatial positioning differences being determined according to spatial positioning differences between said reference and source descriptors.

13. The system of claim 12, wherein said reference descriptors and said source descriptors are local extrema voxels in said reference medical image and said source medical image respectively.

14. The system of claim 9, wherein said registration module comprises an optic flow sub-module, wherein said optic flow sub-module is configured for estimating a spatial optic-flow field according to said spatial positioning differences, said optic-flow field being used during said spatial image registration.

17. The system of claim 14, wherein said segmentation unit is configured for identifying a plurality of source segments in said source medical image, wherein each said source segment matches a respective reference segment of said plurality of reference segments.

18. The system of claim 1, wherein said segmentation unit generates said segmented source medical image by matching voxels of said reference medical image with voxels of said source medical image.

19. The system of claim 17, wherein said source segments are used for identifying said features.

20. The system of claim 17, further comprising an active contour (AC) module, said segmentation unit being configured to forward said plurality of source segments to said AC module, said AC module being configured to refine said source segments according to said source medical image.

23. The system of claim 20, wherein said refinement is performed according to the Hounsfield unit (HU) values of voxels of said source segments.

30. The system of claim 5, wherein at least one of said features is selected from the group comprising features that are:

(a) calculated according to the volume of an area of said body organ in relation to a respective volume of a respective area in said reference normal organ;

(b) calculated according to the center-of-mass of an area of said body organ in relation to a respective center-of-mass in a respective area in said reference normal organ;

(c) calculated according to second order moments of an area of said body organ;

(d) calculated according to the surface area of an area of said body organ in relation to a respective surface area in a respective area in said reference normal organ;

(e) calculated according to the average color of voxels in an area of said body organ;

(f) calculated according to one or more characteristics of the shape of an area of said body organ in relation to a respective one or more characteristics of the shape of a respective area in said reference normal organ;

(g) calculated according to the global motion in an area of said body organ in relation to the positioning of said reference normal organ;

(h) calculated according to hyperdense/hypodense areas in an area of said body organ;

(i) calculated according to the symmetry between at least two areas of said body organ; and

(j) calculated according to a mean color of at least one area in said source medical image.

36. The system of claim 30, wherein said one or more characteristics are estimated according to a spherical Fourier transform.

37. The system of claim 30, wherein said one or more characteristics are estimated according to a predefined curvature.

41. The system of claim 1, wherein said body organ is a depicted human brain which is segmented according to two or more of:

white matter, gray matter lateral ventricles, caudate nucleus, internal capsule, thalamus, lentiform nucleaus, sylvian fissue.

42. The system of claim 4, wherein at least one of said features is calculated according to the uniformity of a skull that being depicted in said source medical image.

43. The system of claim 4, wherein at least one of said features is calculated according to the ratio between the measures of ventricles depicted in said source medical image.

44. The system of claim 4, wherein at least one of said features is calculated according to the volume of a cerebrospinal fluid (CSF) in an upper temple depicted in said source medical image.

45. The system of claim 4, wherein at least one of said features is calculated according a minimum mean square error between said reference and source medical images.

46. The system of claim 4, wherein at least one of said features is calculated according an estimation of cerebrospinal fluid of in a sub-arachnoid'"'"'s space in said depicted human brain.

47. The system of claim 1, wherein at least one of said features is calculated according to the existence of predefined structures in said segmented source medical image, said existence is determined by matching segments of said segmented source medical image with a set of normal segmentation structures.

48. The system of claim 1, wherein at least one of said features is calculated according a distance between segments in said segmented source medical image.

51. The system of claim 4, wherein at least one of said features is calculated according to the positioning of a middle symmetry line that bisects said depicted human brain.

54. The system of claim 51, wherein said symmetry is based on the symmetry level of an MSP.

56. The system of claim 4, wherein at least one of said features is calculated according to a distance between at least two temporal horns depicted in said source medical image.

57. The system of claim 1, comprising a classification unit which further comprises a diagnosis module, said diagnosis module being configured for generating a diagnosis report according to said features and said pathology detector.

64. The system of claim 1, comprising a classification unit comprises at least one classifier, said at least one classifier being configured to classify said source medical image according to said features, said classification being determined according to said at least one classifier.

112. The system according to claim 1, wherein said pathology detector analyses a vector including a plurality of features, and wherein said detector includes a memory storing therein a plurality of matches between vectors and abnormal situations.

113. The system according to claim 1, wherein said pathology detector aggregates a plurality of features into a single aggregate feature.

114. The system according to claim 1, wherein said features includes two or more of, a HU density related property, a statistical scattering property of the HU density, a geometrical characteristic, a symmetry characteristic, a geometrical relation between different structures.

115. The system according to claim 1, wherein said pathology detector includes a collection of features and normative values for each segment of said.

2-3. -3. (canceled)

11. (canceled)

15-16. -16. (canceled)

21-22. -22. (canceled)

24-29. -29. (canceled)

31-35. -35. (canceled)

38-40. -40. (canceled)

49-50. -50. (canceled)

52-53. -53. (canceled)

55. The system of claim, wherein at least one of said features is calculated according to a symmetry value, said symmetry value being calculated by measuring the size of at least one portion of at least one hemisphere depicted in said source medical image.

58-63. -63. (canceled)

65. A method for analyzing a source medical image of a body organ, said method comprising:

a) receiving the source medical image;

b) segmenting said source medical image by matching with a reference medical image of a normal organ already segmented;

c) obtaining information about features in the body organ from said segmented source medical image; and

d) detecting at least one deviation from normality in said organ based on said features.

66-103. -103. (canceled)

104. A system for analyzing a source medical image of a brain, said system comprising:

an input unit for obtaining said source medical image, said source medical image having at least three dimensions;

a segmentation unit, operatively associated with said input unit, configured for segmenting said source medical image into a predefined set of segments;

a memory storing therein, for each segment of said brain, at least one feature; and

a feature extraction unit, operatively associated with said segmentation unit, configured for obtaining said features from said segmented source medical image,

wherein said feature extraction unit extracts for brain parts that appear on both sides of the brain, at least one ratio between feature values of said brain parts.

105. A system according to claim 104, wherein said features include at least one geometrical feature.

106. A system according to claim 104, wherein said features include at least one non-geometrical feature.

107. A system according to claim 104, wherein said features include at least aggregate feature as a vector of features.

108. A system according to claim 104, wherein said memory stores therein at least one normative value for each feature.

109. A system according to claim 104, wherein said memory stores therein at least an indication of a plurality of normative distribution for at least some of said features.

110. A system according to claim 104, wherein said features are different for different segments.

111. A system according to claim 104, comprising a pathology detector configured to indicate a plurality of simultaneous suspected pathologies.