Gel electrophoresis image warping

• US 20020027564A1
• Filed: 09/19/2001
• Published: 03/07/2002
• Est. Priority Date: 10/05/1999
• Status: Abandoned Application

• Current Assignee: John Taylor Jr
• Original Assignee: John Taylor Jr
• Inventors: Taylor, John Jr.
• Application Number:
  US09/955,311
• Publication Number:
  US 20020027564A1
• Time in Patent Office: Days
• US Class Current: 345/646
• CPC Class Codes: G06T 3/153 using elastic snapping
• CPC Class Codes: G06T 5/80 Geometric correction
• CPC Class Codes: G06T 7/33 using feature-based methods
• CPC Class Codes: G06V 10/754 involving a deformation of the sample pattern or of the reference pattern; Elastic matching

Abstract

A computer-implemented image processing method and apparatus for warping a plurality of gel electrophoresis images is provided. The method includes the steps of assigning tiepoints in a reference image and in one or more object images. The tiepoints in the object image are evaluated one-by-one by comparison to regions about a corresponding tiepoint in the reference image, and the location of the tiepoint in the object image is adjusted by slight movement to a location with respect to recognizable features in both the reference and object image thereby defining a tiepoint pair linking a location in the reference image with a location in the object image. Outlier tiepoint pairs may be rejected if that pair does not meet predetermined conditions. Warping functions are generated and then globally optimized. The plurality of images are tied together using the tiepoint pairs such that all of the images may be subsequently warped into registration to a single base image selected from the plurality of images.

Claims

1. A computer implemented image processing method for warping a plurality of homologous gel electrophoresis images, comprising the steps of:

assigning multiple tiepoints in each of the homologous images, each tiepoint connecting a region of one image to an approximate corresponding region of one other image;

adjusting tiepoint locations to maximize similarities between regions;

creating warping functions based upon at least some of the tiepoints; and

warping each of the images using the warping functions to bring the tiepoints of each of the plurality of images into close alignment.

2. The computer-implemented image processing method of claim 1, further comprising a step of rejecting those tiepoints that are outliers.

3. The computer-implemented image processing method of claim 1, further comprising a step of optimizing the warping functions before said warping step.

4. The computer-implemented image processing method of claim 1, further comprising a step of selecting a base image from the plurality of images.

5. The computer-implemented image processing method of claim 4, wherein said plurality of images were taken of one gel electrophoresis sample over time.

6. The computer-implemented image processing method of claim 4, wherein said plurality of images were taken of one gel electrophoresis sample over a plurality of exposures.

7. The computer-implemented image processing method of claim 5, wherein said base image is a middle image in said plurality of gel electrophoresis images with respect to time.

8. The computer-implemented image processing method of claim 6, wherein said base image is a middle image in said plurality of gel electrophoresis images with respect to exposure.

9. The computer-implemented image processing method of claim 1, wherein said plurality of images were taken of a plurality of gel electrophoresis samples.

10. The computer-implemented image processing method of claim 1, wherein said gel electrophoresis is a two-dimensional gel electrophoresis slab.

11. The computer-implemented image processing method of claim 1, wherein adjustment of tiepoints is effected in pairs of images.

12. The computer-implemented image processing method of claim 11, wherein said pairs are non-sequential pairs of images.

13. The computer-implemented image processing method of claim 11, wherein said pairs are sequential pairs of images.

14. The computer-implemented image processing method of claim 1, wherein said plurality of electrophoresis images are obtained at predetermined time intervals.

15. The computer-implemented image processing method of claim 14, wherein said predetermined time intervals are fixed time intervals.

16. The computer-implemented image processing method of claim 14, wherein said predetermined time intervals are randomly determined time intervals.

17. The computer-implemented image processing method of claim 13, wherein said plurality of images are obtained at irregular time intervals.

18. The computer-implemented image processing method of claim 1, wherein each of said images comprises a plurality of spots, each of said spots representing an occurrence of a protein.

19. The computer-implemented image processing method of claim 1, wherein each of said images comprises a plurality of spots, each of said spots representing an occurrence of a nucleic acid.

20. The computer-implemented image processing method of claim 1, wherein said adjusting tiepoints step further includes the steps of:

calculating a correlation number for a plurality of tiepoints for each region of the one image with each corresponding region of the one other image;

choosing a maximum correlation number as a tiepoint location in said object image; and

recording a position of said corresponding region having said maximum correlation number.

21. The computer-implemented image processing method of claim 20, wherein each of said images is defined by a plurality of pixels, and said choosing step further includes the steps of:

comparing said maximum correlation number to a predetermined threshold; and

determining whether cross-correlation numbers for adjacent pixels decrease monotonically away from said maximum correlation number for a predetermined number of pixels.

22. The computer-implemented image processing method of claim 21, wherein said predetermined threshold includes a correlation number ranging from about 0.4 to about 0.8.

23. The computer-implemented image processing method of claim 21, wherein said predetermined threshold is 0.65.

24. The computer-implemented image processing method of claim 21, wherein said cross-correlation numbers for neighboring points decrease monotonically away from said maximum correlation number for 4 pixels.

25. The computer-implemented image processing method of claim 20, wherein the step of calculating said correlation number includes calculating said correlation number using a fast Fourier transform.

26. The computer-implemented image processing method of claim 1, wherein said assigning tiepoints step includes initially assigning a grid of substantially evenly spaced (x,y) coordinates as said tiepoints in said one image.

27. The computer-implemented image processing method of claim 4, wherein said warping function aligns all tiepoints in all other images to corresponding tiepoints in said base image.

28. The computer-implemented image processing method of claim 1, wherein a predicted X coordinate X_P of said warping function is X_P=C₀+(C₁)(X_B)+(C₂)(Y_B)+(C₃)(X_B²)+(C₄)(X_B)(Y_B)+(C₅)(Y_B²), and wherein a predicted Y coordinate Y_P of said warping function is Y_P=D₀+(D₁)(X_B)+(D₂)(Y_B)+(D₃)(X_B²)+(D₄)(X_B)(Y_B)+(D₅)(Y_B²), where C₀ . . . C₅ and D₀ . . . D₅ are optimized coefficients.

29. The computer-implemented image processing method of claim 1, wherein the warping step includes an interpolation of a warped pixel.

30. The computer-implemented image processing method of claim 4, wherein an image intensity value of a warped pixel is distributed among one or more target pixels in order to warp said warped pixel to said base image.

31. The computer-implemented image processing method of claim 1, wherein the warping step interpolates a warped pixel P_W into four adjacent pixels a, b, c, and d having intensity values V_a, V_b, V_c, and V_d and having respective overlapping regions A_a, A_b, A_c, and A_d according to the formula P_W=(A_a)(V_a)+(A_b)(V_b)+(A_c)(V_c)+(A_d)(V_d).

32. The computer-implemented image processing method of claim 2, wherein said rejecting step further includes the steps of:

calculating functions to predict object image tiepoints from reference image tie points;

calculating distances between predicted object image tiepoints and actual object image tiepoints;

calculating standard deviation of said distances; and

rejecting object image tiepoint as an outlier if said distance fails to meet rejection criteria.

33. The computer-implemented image processing method of claim 3, wherein said optimizing step further includes the steps of:

creating inverse warping functions;

projecting object image tiepoints to the base image;

projecting reference image tiepoints to the base image;

calculating distances between the object image tiepoint projections and reference image tiepoint projections;

calculating standard deviation of said distance; and

rejecting object image tiepoints as outliers if said distances fails to meet rejection criteria.

35. The computer-implemented image processing apparatus of claim 34, wherein said computer is configured to select a base image and warp said plurality of gel electrophoresis images into alignment with said base image.

36. The computer-implemented image processing apparatus of claim 34, wherein said image capturing device is a digital camera.

37. The computer-implemented image processing apparatus of claim 34, wherein said image capturing device is a video camera.

38. The computer-implemented image processing apparatus of claim 34, wherein said image capturing device is a scanner.

39. The computer-implemented image processing apparatus of claim 34, wherein said plurality of gel electrophoresis images are stored in said memory.

40. The computer-implemented image processing apparatus of claim 34, further including an output device for outputting a plurality of warped gel electrophoresis images.

41. The computer-implemented image processing apparatus of claim 34, further including a printer.

42. The computer-implemented image processing apparatus of claim 34, further including a display.

43. The computer-implemented image processing apparatus of claim 34, further including a communications link connected to said computer and over which a plurality of warped gel electrophoresis images may be transferred to a remote location.

44. The computer-implemented image processing apparatus of claim 34, is configured to assign multiple tiepoints in each of the homologous images, each tiepoint connecting a region of one image to an approximate corresponding region of one other image, adjust tiepoint locations to maximize similarities between regions, create warping functions based upon at least some of the tiepoints, and warp each of the images using the warping functions to bring the tiepoints of each of the plurality of images into close alignment.

45. The computer-implemented image processing apparatus of claim 44, is further configured to reject those tiepoints that are outliers.

46. The computer-implemented image processing apparatus of claim 44, is further configured optimize the warping functions before said warping step.

47. The computer-implemented image processing apparatus of claim 46, wherein said plurality of images were taken of one gel electrophoresis sample over time.

48. The computer-implemented image processing apparatus of claim 47, wherein said gel electrophoresis is a two-dimensional gel electrophoresis slab.

49. The computer-implemented image processing apparatus of claim 45, wherein said apparatus is further configured to calculate functions to predict object image tiepoints from reference image tie points, calculate distances between predicted object image tiepoints and actual object image tiepoints, calculate standard deviation of said distances, and reject object image tiepoints as outliers if said distances fails to meet rejection criteria.

50. The computer-implemented image processing apparatus of claim 45, wherein said apparatus is further configured to create inverse warping functions, project object image tiepoints to the base image, project reference image tiepoints to the base image, calculate distances between the object image tiepoint projections and reference image tiepoint projections, calculate standard deviation of said distance, and reject object image tiepoints as outliers if said distances fails to meet rejection criteria.

34. A computer-implemented image processing apparatus for warping a plurality of gel electrophoresis images, comprising:

an image capturing device; and

a computer having a memory and configured to communicate with said image capturing device, said computer capable of receiving and storing into said memory a plurality of gel electrophoresis images from said image capturing device, said computer being further configured for warping said plurality of gel electrophoresis images.