Variable-length correlation method for motion correction in SPECT myocardial perfusion imaging

• US 6,473,636 B1
• Filed: 02/03/2000
• Issued: 10/29/2002
• Est. Priority Date: 02/03/2000
• Status: Expired due to Term

• Current Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Original Assignee: Siemens Corporate Research Incorporated (Siemens AG)
• Inventors: Chen, Eric Q., Wei, Guo-Qing, Qian, Jianzhong, Engdahl, John C.
• Primary Examiner(s): Lateef, Marvin M.
• Assistant Examiner(s): MANTIS MERCADER, ELENI M
• Application Number:
  US09/496,764
• Time in Patent Office: 999 Days
• Field of Search: 600/436, 600/407, 600/425, 250/363.02, 382/128, 128/922
• US Class Current: 600/436
• CPC Class Codes: A61B 5/0275 using tracers, e.g. dye dilution
• CPC Class Codes: A61B 5/416 the spleen
• CPC Class Codes: A61B 5/4244 liver
• CPC Class Codes: Y10S 128/922 including image analysis

Abstract

According to one embodiment of the invention, a variable length correlation method is provided for compensating for body motions and heart creep in a sequence of image frames obtained by single-photon emission computed tomograph (SPECT) myocardial perfusion imaging. Abrupt body motions, gradual body motions, and heart upward creep are detected and corrected in the sequence of image frames, based on varying correlation lengths. A linear transformation is used to compensate for appearance changes of the heart under different acquisition angles so as to achieve high accuracy of motion detection and correction. The detecting and correcting step is applied in the two-dimensional image domain. The method also includes the steps of detecting and excluding high intensity, non-heart regions in a particular image frame so that non-heart motions such as bowel gas, spleen, and liver motions are excluded from being involved in the motion detection and correction procedure.

Claims

1. A variable length correlation method for compensating for body motions and heart creep in a sequence of image frames obtained by single-photon emission computed tomograph (SPECT) myocardial perfusion imaging, comprising the steps of:

determining whether a bright spot corresponding to a non-heart motion exists in a particular image frame;

automatically setting a y-coordinate lower limit of the particular image frame as an offset from a position of the bright spot, when the bright spot exists;

automatically setting the y-coordinate lower limit of the particular image as an offset from a bottom of the particular image frame, when the bright spot does not exist; and

detecting and correcting abrupt body motions, gradual body motions, and heart upward creep in the sequence of image frames, based on varying correlation lengths, wherein said said detecting and correcting step is applied in the two-dimensional image domain.

2. The method according to claim 1, wherein said determining step comprises the step of determining consistency of sizes and positions of binarized image blobs using multiple thresholds.

3. The method according to claim 1, wherein said determining step comprises the steps of:

binarizing a given image frame by successively lowering a threshold value;

identifying connected regions in the given image;

determining pre-identified characteristics of each of the connected regions; and

identifying a given connected region as the bright spot, when the pre-identified characteristics of the given connected region corresponds to predetermined criteria.

4. The method according to claim 3, wherein the pre-identified characteristics comprise at least some of an area, a center of gravity, an x-size, and a y-size.

5. The method according to claim 3, wherein the predetermined criteria comprise at least some of an x and y size-ratio being within a first pre-specified limit, an area being within a second pre-specified limit, and continuous changes across the multiple thresholds.

6. The method according to claim 1, wherein said detecting and correcting step comprises the step of comparing a given image frame to earlier image frames of varying distances from the given image frame.

7. The method according to claim 1, wherein said detecting and correcting step comprises the step of comparing a given image frame to at least one earlier image frame having a predefined distance from the given image frame, for each of the abrupt body motions, the gradual body motions, and the heart upward creep, respectively.

8. The method according to claim 7, wherein said detecting and correcting step comprises the step of applying a gradient descent method and a bilinear function interpolation to minimize a correlation function corresponding to the given image frame and the at least one image frame, so as to achieve sub-pixel accuracy of motion estimation.

9. The method according to claim 1, wherein the abrupt body motions, the gradual body motions, and the heart upward creep are each determined and corrected during a separate pass of said method, wherein each of the separate passes comprises the step of comparing a given image frame to at least one earlier image frame having a predefined distance from the given image frame.

10. The method according to claim 1, wherein a first pass detects the abrupt body motions, a second pass detects the gradual body motions, and a third pass detects the heart upward creep.

11. The method according to claim 1, wherein said detecting and correcting step comprises the step of applying a linear transformation to at least some of the image frames to compensate for the x-directional appearance changes caused by viewing angle differences.

12. The method according to claim 11, wherein said applying step comprises the step of applying the linear transformation to one of every two image frames.

13. A variable length correlation method for compensating for body motions and heart creep in a sequence of image frames obtained by single-photon emission computed tomograph (SPECT) myocardial perfusion imaging, comprising the steps of:

automatically setting a y-coordinate lower limit of a particular image frame as an offset from a position of a bright spot corresponding to a non-heart motion, when the bright spot exists in the particular image frame; and

detecting and correcting abrupt body motions, gradual body motions, and heart upward creep in the sequence of image frames, wherein said detecting and correcting step is applied in the two-dimensional image domain, and includes the step of comparing a given image frame to at least one earlier image frame having a predefined distance from the given image frame, for each of the abrupt body motions, the gradual body motions, and the heart upward creep, respectively.

14. A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for compensating for body motions and heart creep in a sequence of image frames obtained by single-photon emission computed tomograph (SPECT) myocardial perfusion imaging, the method based on variable length correlation, and comprising the steps of:

determining whether a bright spot corresponding to a non-heart motion exists in a particular image frame;

automatically setting a y-coordinate lower limit of the particular image frame as an offset from a position of the bright spot, when the bright spot exists;

automatically setting the y-coordinate lower limit of the particular image as an offset from a bottom of the particular image frame, when the bright spot does not exist; and

detecting and correcting abrupt body motions, gradual body motions, and heart upward creep in the sequence of image frames, based on varying correlation lengths, wherein said detecting and correcting step is applied in the two-dimensional image domain.

15. The program storage device according to claim 14, wherein said determining step comprises the step of determining consistency of sizes and positions of binarized image blobs using multiple thresholds.

16. The program storage device according to claim 14, wherein said determining step comprises the steps of:

binarizing a given image frame by successively lowering a threshold value;

identifying connected regions in the given image;

determining pre-identified characteristics of each of the connected regions; and

identifying a given connected region as the bright spot, when the pre-identified characteristics of the given connected region corresponds to predetermined criteria.

17. The program storage device according to claim 16, wherein the pre-identified characteristics comprise at least some of an area, a center of gravity, an x-size, and a y-size.

18. The program storage device according to claim 16, wherein the predetermined criteria comprise at least some of an x and y size-ratio being within a first pre-specified limit, an area being within a second pre-specified limit, and continuous changes across the multiple thresholds.

19. The program storage device according to claim 14, wherein said detecting and correcting step comprises the step of comparing a given image frame to earlier image frames of varying distances from the given image frame.

20. The program storage device according to claim 14, wherein said detecting and correcting step comprises the step of comparing a given image frame to at least one earlier image frame having a predefined distance from the given image frame, for each of the abrupt body motions, the gradual body motions, and the heart upward creep, respectively.

21. The program storage device according to claim 20, wherein said detecting and correcting step comprises the step of applying a gradient descent method and a bilinear function interpolation to minimize a correlation function corresponding to the given image frame and the at least one earlier image frame, so as to achieve sub-pixel accuracy of motion estimation.

22. The program storage device according to claim 14, wherein the abrupt body motions, the gradual body motions, and the heart upward creep are each determined and corrected during a separate pass of said method, wherein each of the separate passes comprises the step of comparing a given image frame to at least one earlier image frame having a predefined distance from the given image frame.

23. The program storage device according to claim 14, wherein a first pass detects the abrupt body motions, a second pass detects the gradual body motions, and a third pass detects the heart upward creep.

24. The program storage device according to claim 14, wherein said detecting and correcting step comprises the step of applying a linear transformation to at least some of the image frames to compensate for the x-directional appearance changes caused by viewing angle differences.