Method for correction of relative object-detector motion between successive views

• US 20050010108A1
• Filed: 06/24/2004
• Published: 01/13/2005
• Est. Priority Date: 04/19/2002
• Status: Active Grant

• Current Assignee: VisionGate, Inc.
• Original Assignee: VisionGate, Inc.
• Inventors: Nelson, Alan C., Rahn, John Richard
• Granted Patent: US 7,260,253 B2
• Time in Patent Office: Days
• US Class Current: 600/425
• CPC Class Codes: G01N 15/1433 using image recognition
• CPC Class Codes: G01N 15/1468 with spatial resolution of the texture or inner structure of the particle
• CPC Class Codes: G01N 15/147 the analysis being performed on a sample stream
• CPC Class Codes: G01N 2015/1006 for cytology
• CPC Class Codes: G01N 2015/1027 Determining speed or velocity of a particle
• CPC Class Codes: G01N 2021/0346 Capillary cells; Microcells
• CPC Class Codes: G01N 2021/1787 Tomographic, i.e. computerised reconstruction from projective measurements
• CPC Class Codes: G01N 21/05 Flow-through cuvettes G01N21/09 takes precedence; handling fluid samples G01N1/10
• CPC Class Codes: G01N 21/4795 spatially resolved investigating of object in scattering medium in vivo A61B
• CPC Class Codes: G01N 2201/0813 Arrangement of collimator tubes, glass or empty
• CPC Class Codes: G01N 2223/612 biological material
• CPC Class Codes: G02B 21/26 Stages; Adjusting means therefor
• CPC Class Codes: G02B 21/365 Control or image processing arrangements for digital or video microscopes G02B21/361, G02B21/362 take precedence
• CPC Class Codes: G06T 11/005 Specific pre-processing for tomographic reconstruction, e.g. calibration, source positioning, rebinning, scatter correction, retrospective gating
• CPC Class Codes: G06T 11/006 Inverse problem, transformation from projection-space into object-space, e.g. transform methods, back-projection, algebraic methods
• CPC Class Codes: G06T 2211/412 Dynamic
• CPC Class Codes: G06T 2211/421 Filtered back projection [FBP]

Abstract

Registration correction for optical tomographic imaging in three dimensions. An object of interest is illuminated to produce an image. A lateral offset correction value is determined for the image. An axial offset correction value is determined for the image. The lateral offset correction value and the axial offset correction value are applied to the image to produce a corrected file image.

Claims

1. A method for correction of relative object-detector motion between successive views comprising the steps of:

illuminating an object of interest to produce an image;

determining a lateral offset correction value for the image;

determining an axial offset correction value for the image; and

applying the lateral offset correction value and the axial offset correction value to the image to produce a corrected file image.

2. The method of claim 1, wherein the step of determining a lateral offset correction value for the image further comprises the steps of:

thresholding the image; and

cross-correlating the image with a template image.

3. The method of claim 2, wherein the template image is created by a method comprising the steps of:

creating at least two white lines at predetermined positions to form a preliminary template image;

expanding the preliminary template image to provide zero-padding in two dimensions resulting in an expanded template image; and

performing a two-dimensional FFT on the expanded template image to create the final template image.

4. The method of claim 2, wherein the steps of thresholding the image and cross-correlating the image further comprise the steps of:

finding the grayscale histogram of the image including a plurality of bins;

identifying a bin with the greatest number of pixels;

setting all pixels in the image having that bin'"'"'s grayscale value or higher equal to zero in a copy of the image;

applying a two-dimensional FFT to the copy of the image to produce a Fourier transform;

multiplying the Fourier transform by a complex conjugate of the Fourier transform of the template image to produce a new image array;

summing the new image array, along each of the plurality of rows to compute a lateral sum array;

computing a one dimensional Fourier transform of the lateral sum array to find the cross-correlation of the rows of the copy of the original image and the template image;

setting an uncorrected position of the upper tube wall, at the location of the maximum value of the lateral sum array; and

determining the lateral offset as the difference between the uncorrected position of the upper tube wall and a predetermined wall edge position.

5. The method of claim 1, wherein the step of determining an axial offset correction value for the image further comprises the steps of:

thresholding the image;

cross-correlating the image with the thresholded version of a previous image; and

determining an axial offset as a maximum in the cross-correlation function along the line that corresponds to the difference in the lateral corrections of the two images.

6. The method of claim 1, wherein the step of determining an axial offset further comprises the steps of:

finding the grayscale histogram of the image including a plurality of bins;

identifying a bin with the greatest number of pixels;

setting all pixels in the original image having that bin'"'"'s grayscale value or higher equal to zero in a copy of the original image;

applying a low-pass filter to the thresholded image;

computing a cross-correlation of the thresholded, low-pass filtered image with a preceding image'"'"'s thresholded, low-pass filtered version;

finding the maximum correction value in the row of the resultant cross-correlation that corresponds to the difference in the two images'"'"' lateral offsets; and

adding the correction value to the sum of all the previous axial offsets.

7. The method of claim 1 further comprising the steps of:

writing the value of the lateral offset to a file;

writing the value of the array element used to determine the lateral offset to a file;

writing the value of the axial offset to a file;

writing the value of the array element used to determine the axial offset to a file; and

generating a corrected image by cropping the appropriate number of pixels from one or two edges and shifting the remaining pixels by the number cropped.

8. The method of claim 1, wherein the object of interest is a cell or a cell nucleus.

9. The method of claim 4, further comprising the step of repeating the steps with a different number of bins in the histogram, if the maximum value of the cross-correlation has a magnitude less than a predetermined value.

22. The method of claim 1 wherein a plurality of images are produced to generate an input image file and a plurality of corrected file images are a cropped copy of the input image file, with uncropped portions shifted vertically and/or horizontally, and with additional blank pixels inserted at one or two of the borders to retain the input file image size.

23. The method of claim 1 wherein the corrected file images comprise calculated offset values saved to a digital file.

24. The method of claim 23 wherein the calculated offset values are processed using computer word-processing to produce an altered text file, and the altered text file is used to generate offsets in two axes, and wherein the step for determining a lateral offset correction value for the image is iterated to find a maximum value.

25. The method of claim 22 wherein the calculated offset values comprise:

a position of a maximum value of the cross-correlation between a current image and the template image;

a maximum value of the cross-correlation between a current image and the template image;

a location of a maximum correlation value on the corresponding row of a cross-correlation between the current image and the previous image; and

a maximum value of the cross-correlation between a current image and a preceding image.

10. Apparatus for correcting positioning and motion errors in an imaging system, comprising:

means for acquiring at least two views of an object;

said at least two views including M rows and N columns of image brightness data;

means for comparing a pattern of image brightness data from each acquired view with a corresponding reference pattern of image brightness data to produce two coefficients “

A” and

“

B”

containing information about a similarity therebetween;

said means for comparing includes;

first means for producing a modified copy of said acquired image in which an alteration of brightness level is applied to pixels having brightness levels within a specified range;

second means for performing a cross-correlation between said modified copy of the acquired view with said reference view;

third means for locating and evaluating the maximum of the resulting cross-correlation function, the row or column where said maximum is located determining the value of coefficient “

A”

;

fourth means for producing modified copies of said acquired images in which an alteration of brightness level is applied to pixels having brightness levels within a specified range;

fifth means for performing a cross-correlation between two said modified copies of consecutive acquired images;

sixth means for locating and evaluating the restricted maximum of the resulting cross-correlation function, said restriction being that the row or column on which said restricted maximum is located must be orthogonal to and intersect the row or column on which coefficient “

A”

is located;

location of said restricted maximum determining coefficient “

B.”

11. A method according to claim 10, further comprising means responsive to said values of said coefficients “

A” and

“

B”

for shifting the acquired views in two orthogonal directions, the directions and magnitudes of said amounts being related to said values of said coefficients “

A” and

“

B.”

12. A method according to claim 10, in which the means for cross-correlating to determine coefficient “

A”

includes the step of summing the rows or columns to obtain a one-dimensional array, said summation being performed in a direction orthogonal to the direction of the shift whose magnitude is determined by said coefficient “

A.”

13. A method according to claim 10 in which the reference pattern comprises an array of M rows and N columns of image brightness data in which all pixels, with the exception of one or more rows or columns, have the same brightness value, said one or more rows or columns having a brightness that is uniform, but substantially different than the other pixels.

14. A method according to claim 10 in which the reference pattern comprises an array of M rows and N columns of image brightness data in which all pixels, with the exception of two groups consisting of one or more rows or columns, have the same brightness value, said two groups having a brightness that is uniform, but substantially different than the other pixels, and having a separation that corresponds to the apparent separation of the opposing walls of a microcapillary tube, said opposing walls both appearing in the field of view of a succession of two or more images, said microcapillary tube containing an object of interest.

15. A method according to claim 10 in which the object of comprises a biological cell or nucleus.

16. A method according to claim 10 including means of repeating the steps of thresholding, cross-correlation, and locating and evaluating the maximum of said cross-correlation, said repetition comprising means of finding the maximum value of the maximum correlation based on multiple thresholding.

17. A method according to claim 14 in which the separation of the two groups or rows or columns is computed based on one or more acquired images.

18. A method for three dimensional (3D) reconstruction of an object of interest, comprising the steps of:

(a) packing a set of objects of interest into a linear container;

(b) illuminating at least at least one object of the set of objects of interest with at least one optical projection beam;

(c) translating the linear container until the at least one object is located within a region of the at least one optical projection beam;

(d) centering the at least one object as necessary;

(e) rotating the at least one object through a plurality of radial angles;

(f) generating a set of projection images at each radial angle of the plurality of angles;

(g) repeating the steps (b) through (f) until the set of objects of interest (1) has been scanned; and

(h) correcting motion of the scanned objects of interest by determining a lateral offset correction value for the scanned image, determining an axial offset correction value for the scanned image, and applying the lateral offset correction value and the axial offset correction value to the scanned image to produce a corrected file image.

19. The method of claim 18, wherein the at least one object is a cell or a cell nucleus.

20. The method of claim 18, wherein the step of packing a set of objects of interest into a linear container comprises the steps packing a plurality of cells into a microcapillary tube.

21. A method for correction of relative object-detector motion between successive views comprising the steps of:

imaging an object of interest to produce an image;

determining an axial offset correction value for the image, wherein the step of determining an axial offset correction value for the image further includes the steps of;

thresholding the image, cross-correlating the image with the thresholded version of a previous image, and determining an axial offset as a maximum in the cross-correlation function along the line that corresponds to the difference in the lateral corrections of the two images; and

applying the lateral offset correction value and the axial offset correction value to the image to produce a corrected file image.