System and method for mining quantitive information from medical images

• US 7,158,692 B2
• Filed: 10/15/2002
• Issued: 01/02/2007
• Est. Priority Date: 10/15/2001
• Status: Active Grant

• Current Assignee: Cloud Software Group
• Original Assignee: Insightful Corporation (Cloud Software Group)
• Inventors: Chalana, Vikram, Fogarasi, Stephen, Pathak, Sayan, Racki, Steven, Ng, Lydia, Wyman, Bradley, Sparks, Bobbi, Oelund, John
• Primary Examiner(s): PATEL, KANJIBHAI B
• Application Number:
  US10/271,916
• Publication Number:
  US 20030174872A1
• Time in Patent Office: 1,540 Days
• Field of Search: 382/173, 382/289, 382/305, 382/190, 382/181, 382128-134, 382/294, 600/410, 707/104
• US Class Current: 382/294
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images
• CPC Class Codes: G16H 50/70 for mining of medical data, e.g. analysing previous cases of other patients
• CPC Class Codes: G16Z 99/00 Subject matter not provided for in other main groups of this subclass

Abstract

A system for image registration and quantitative feature extraction for multiple image sets. The system includes an imaging workstation having a data processor and memory in communication with a database server in communication. The a data processor capable of inputting and outputting data and instructions to peripheral devices and operating pursuant to a software product and accepts instructions from a graphical user interface capable of interfacing with and navigating the imaging software product. The imaging software product is capable of instructing the data processor, to register images, segment images and to extract features from images and provide instructions to store and retrieve one or more registered images, segmented images, quantitative image features and quantitative image data and from the database server.

Claims

1. A system for image registration and quantitative feature extraction from the images, the system comprising:

an imaging workstation; and

a database server in communication with the imaging workstation, wherein the imaging workstation includes;

a data processor with memory capable of inputting and outputting data and instructions to peripheral devices;

a graphical user interface capable of interfacing with and navigating an imaging software product, wherein the imaging software product is capable of instructing the data processor, the imaging software product comprises;

instructions to register images;

instructions to segment images;

instructions to extract features from images; and

instructions to store and retrieve one or more registered images, segmented images, quantitative image features and quantitative image data from the database server.

2. The system of claim 1, wherein the software product further includes instructions to perform queries on the database server.

3. The system of claim 1, wherein the software product further comprises instructions to perform multi-modality image registration.

4. The system of claim 1, wherein the software product further includes instructions to perform multi-channel region segmentation.

5. The system of claim 1, wherein the software product further includes instructions to perform shape-based interpolation.

6. The system of claim 1, wherein the quantitative image data further includes serial/longitudinal data.

7. The system of claim 6, wherein the quantitative image data is extracted from 2D and 3D images.

8. The system of claim 1, wherein the quantitative image data is extracted from 2D and 3D images.

9. The system of claim 5, wherein the shaped-based interpolation is modified to perform a regional adjustment of the contour based on the image properties.

10. The system of claim 1, wherein the software product further includes instructions to fuse information from multiple modalities.

11. The system of claim 1, wherein the software product further includes instructions to display segmented images as one of translucent regions and boundaries as an overlay on top of an image.

12. The system of claim 11, wherein the degree of translucency of the regions and boundaries is adjustable.

13. The system of claim 1, that further includes capabilities for generating and storing audit trails.

14. The system of claim 1, wherein the software product further comprises instructions to:

select an initial contour of a structure in an image set;

identify each vertex of an initial contour;

identity radial lines normal to each vertex of the initial contour, wherein the initial contour is restricted to move only along the radial lines;

sample the image set along the radial lines;

compute one-dimensional derivatives of the structure along the radial lines; and

iteratively adjust positioning of the initial contour on the radial line and calculating total energy, wherein the initial contour is adjusted such that the total energy is minimized.

15. The system of claim 14, wherein the total energy is defined by

E(ν

)=Σ

(α

∥

ν

_S∥

²+β

ν

₃₈∥

²+λ

g(ν

)).

16. The system of claim 14, wherein a final radial representation of the contour is convened back to a Cartesian-coordinate for presentation to the user.

17. The system of claim 9, wherein the regional adjustment of the contour based on the local image properties further comprises select an initial contour of a structure in an image set, identify each vertex of an initial contour, identify radial lines normal to each vertex of the initial contour, wherein the initial contour is restricted to move only along the radial lines, sample the image set along the radial lines, compute one-dimensional derivatives of the structure along the radial lines, and iteratively adjust positioning of the initial contour on the radial line and calculating total energy, wherein the initial contour is adjusted such that the total energy is minimized.

18. The system of claim 1, wherein the graphical user interface permits viewing of an arbitrary number of images in a grid by clicking and dragging a grid layout tool that shows a grid layout of the images to be displayed.

19. A computer implemented method for performing a study on image data using an image analysis system having an imaging workstation, a database server in communication with the imaging workstation, wherein the imaging workstation includes a data processor with memory capable of inputting and outputting data and instructions to peripheral devices and a graphical user interface capable of interfacing with and navigating an imaging software product, wherein the imaging software product is capable of instructing the data processor, the method comprising:

importing one or more image sequences from the database server for a one or more patients involved in the study;

registering the images to align the different image sequences with one of each other and to a baseline image set;

performing segmentation to find one or more objects of interest in the entire image sequence;

performing feature extraction to determine quantitative information about the objects of interest; and

storing the objects of interest and the quantitative information about the objects of interest in the database server.

20. The method of claim 19 further comprises performing a regional adjustment of the contour based on the local image characteristics.

21. The method of claim 19, wherein the quantitative image data further includes serial/longitudinal data.

22. The method of claim 21, wherein the quantitative image data is extracted from 2D and 3D images.

23. The method of claim 19, wherein the quantitative image data is extracted from 2D and 3D images.

24. The method of claim 19 further comprising fusing information from multiple modalities.

25. The method of claim 19 further comprising displaying segmented images as one of translucent regions and boundaries as an overlay on top of an image.

26. The method of claim 25 further comprising the ability to vary the translucency of the regions and boundaries between a range or subset of a range from opaque to transparent.

27. The method of claim 20, wherein the regional adjustment further comprising:

selecting an initial contour of a structure in an image set;

identifying each vertex of the initial contour;

identifying radial lines normal to each vertex of the initial contour, wherein the initial contour is restricted to move only along the radial lines;

sampling the image set along the radial lines;

computing one-dimensional derivatives of the structure along the radial lines; and

iteratively adjusting positioning of the initial contour on the radial line and calculating total energy, wherein the initial contour is adjusted such that the total energy is minimized.

28. The method of claim 27, wherein the total energy is defined by

E(ν

)=Σ

(α

∥

ν

_S∥

²+β

∥

ν

_SS∥

²+λ

g(ν

)).

29. The method of claim 27, wherein a final radial representation of the contour is converted back to a Cartesian-coordinate for presentation to the user.

30. The method of claim 27, wherein selecting the initial contour comprises one of the following:

manual painting, polygon drawing, seed-fill, live-wire, multi-channel segmentation and single channel segmentation.

31. The system of claim 19 further comprising storing audit data associated with the images in the database server.

32. An image analysis system, comprising:

an imaging workstation;

a database server in communication with the imaging workstation;

wherein the imaging workstation includes a data processor with memory capable of inputting and outputting data and instructions to peripheral devices and a graphical user interface capable of interfacing with and navigating an imaging software product, wherein the imaging software product is capable of instructing the data processor, the imaging software product comprises;

instructions to store and retrieve one or more registered images, segmented images, quantitative image features and quantitative image data and from the database server; and

instructions to store audit data in the database server that tacks the use and changes of the registered images, the segmented images, the quantitative image features and quantitative image data.

33. A method for segmenting images from a sequence of images, comprising:

performing seeded region growing to generate a segmented region on one or more images and one or more regions of interest, the seeded region growing further comprising identifying a point within a region to be segmented on a particular image and identifying pixels surrounding the point having a similar intensity range; and

performing shaped based interpolation that further comprises identifying the segmented region(s) identified during the seeded region growing on the other images in the sequence of images to segment the identified region(s) in all of the sequence of images.

34. A method for segmenting images from a sequence of images, comprising:

determining an initial contour for a segmented region in an image;

performing shaped based interpolation that further comprises identifying the segmented region identified from the image for the other images in the sequence of images to segment the segmented region in all of the sequence of images; and

performing snake segmentation on segmented regions in the sequence of images to generate a refined segmented region for the sequence of images, the snake segmentation further comprising selecting an initial contour of a structure in an image set;

identifying each vertex of the initial contour;

identifying radial lines normal to each vertex of the initial contour, wherein the initial contour is restricted to move only along the radial lines;

sampling the image set along the radial lines;

computing one-dimensional derivatives of the structure along the radial lines; and

iteratively adjusting positioning of the initial contour on the radial line and calculating total energy, wherein the initial contour is adjusted such that the total energy is minimized.

35. The method of claim 34, wherein the total energy is defined by

E(ν

)=Σ

(α

∥

ν

_S∥

²+β

∥

ν

_SS∥

²+λ

g(ν

)).

36. The method of claim 34, wherein a final radial representation of the contour is converted back to a Cartesian-coordinate for presentation to the user.

37. The method of claim 34, wherein determining the initial contour is one of the following:

manual painting, polygon drawing, seed-fill, live-wire, multi- or single-channel segmentation.

38. The method of claim 34, wherein the method of identifying the initial contour is from using the contours identified on another image series for the same patient or subject.

39. The method of claim 34, wherein the method of identifying the initial contour is from a template or reference image.

40. A method for segmenting images from a sequence of images, comprising:

generating a segmented image from a reference image;

determining registration between the reference image and a new image to generate a new image segmentation;

performing snake segmentation on the new image segmentation to generate a segmented region within the particular image, the snake segmentation further comprising selecting an initial contour of a structure in an image set;

identifying each vertex of the initial contour;

identifying radial lines normal to each vertex of the initial contour, wherein the initial contour is restricted to move only along the radial lines;

sampling the image set along the radial lines;

computing one-dimensional derivatives of the structure along the radial lines; and

iteratively adjusting positioning of the initial contour on the radial line and calculating total energy, wherein the initial contour is adjusted such that the total energy is minimized.

41. A method for segmenting images from a sequence of images, comprising:

performing seed region segmentation to generate a segmented region within the particular image, the seed region segmentation further comprising identifying a point within a region to be segmented on a particular image and identifying pixels surrounding the point having a predetermined intensity; and

performing multi-channel segmentation in order to segment the segmented region from each image in an image set.

42. A method for segmenting images from a sequence of images, comprising:

determining an initial contour for a segmented region in an image, the initial contour being determined a segmentation tool; and

performing snake segmentation on segmented region of the image to generate a refined segmented region for the image, the snake segmentation further comprising selecting an initial contour of a structure in an image set;

identifying each vertex of the initial contour;

identifying radial lines normal to each vertex of the initial contour, wherein the initial contour is restricted to move only along the radial lines;

sampling the image set along the radial lines;

computing one-dimensional derivatives of the structure along the radial lines; and

iteratively adjusting positioning of the initial contour on the radial line and calculating total energy, wherein the initial contour is adjusted such that the total energy is minimized.

43. The method of claim 42, wherein the total energy is defined by

E(ν

)=Σ

(α

∥

ν

_S∥

²+β

∥

v_SS∥

²+λ

g(ν

)).

44. The method of claim 43, wherein a final radial representation of the contour is converted back to a Cartesian-coordinate for presentation to the user.

45. The method of claim 42, wherein the segmentation tool comprises one or more of manual painting, polygon drawing, seed-fill, live-wire, multi-channel segmentation and single-channel segmentation.

46. The method of claim 42, wherein the method of identifying the initial contour is from using the contours identified on another image series for the same patient or subject.

47. The method of claim 42, wherein the method of identifying the initial contour is from a template or reference image.