Optoelectronic system for implementation of iterative computer tomography algorithms

• US 5,414,623 A
• Filed: 05/08/1992
• Issued: 05/09/1995
• Est. Priority Date: 05/08/1992
• Status: Expired due to Fees

• Current Assignee: Iowa State University Research Foundation Incorporated (Iowa State University)
• Original Assignee: Iowa State University Research Foundation Incorporated (Iowa State University)
• Inventors: Lu, Tongxin, Udpa, Satish S.
• Primary Examiner(s): Hayes, Gail O.
• Assistant Examiner(s): TKACS, STEPHEN R
• Application Number:
  US07/880,716
• Time in Patent Office: 1,096 Days
• Field of Search: 364/413.13, 364/413.14, 364/413.16, 364/413.19, 364/413.21, 378/98, 378/99
• US Class Current: 382/131
• 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 2211/424 Iterative

Abstract

An optoelectronic tomographic reconstruction system utilizes spacial light modulators in charge couple devices to perform projection iterative reconstruction techniques and simultaneous iterative reconstruction techniques. A back projection processor uses a linear array of analog spacial light modulators in a cylindrical lens, an image detection array, and an image rotator. The image rotator smears a projected image at the same angle as the projection was taken. The back projection processor thus smears the projection back to the image space. An optoelectronic forward projection processor uses an spacial light modulator array, an image rotator, and an image detecting array. A reconstructed image displayed by the spacial light modulator is smeared by the rotator to forward project the reconstructed image on the image detecting array at the same angle as when the measured projection was taken. The forward projection processor thus smears the reconstructed image back into the projection space.

Claims

1. A system for reconstruction of a two dimensional (2-D) image based on a plurality of one dimensional (1-D) image projections of an object obtained from an imaging device, the projections taken at different projection angles, each 1-D image projection represented by a corresponding image space data vector, each image space data vector associated with a unique respective angle of projection, comprising:

(a) a back projection processor including;

(i) a 2-D charge coupled device (CCD) detecting array,

(ii) first spatial light modulator (SLM) projection means for generating, using spatial light modulation, in response to an input image space data vector representing a 1-D image, a corresponding 1-D image stretched along one axis to provide a stretched 1-D image, and for projecting the stretched 1-D image onto the 2-D CCD detecting array,

(iii) first image rotation means for orienting a stretched 1-D image projected by the SLM projection means on the 2-D CCD detecting array at a desired projection angle, the 2-D CCD detecting array outputting back projected data;

so that the back projection processor smears each 1-D image projection represented in a corresponding image space data vector onto the 2-D CCD detecting array for each respective angle of projection;

(b) a forward projection processor including;

(i) a CCD detecting array,

(ii) second SLM projection means for generating, using spatial light modulation, in response to an input data array representing a 2-D image, a corresponding image compressed along one axis to provide a compressed 1-D image, and for projecting the compressed 1-D image on the CCD detecting array, the CCD detecting array outputting forward projected data,

(iii) second image rotation means for orienting a compressed 1-D image projected by the second SLM projection means on the CCD detecting array at a desired projection angle, so that the forward projection processor forward projects a 2-D image represented in the input data array onto the CCD detecting array for each unique respective angle of projection to generate forward projected vectors corresponding to each respective image space data vector; and

(c) control means connected to said forward projection processor and said back projection processor for performing the following steps;

(i) controlling said back projection processor to perform back projection operations for each respective angle of projection;

(ii) controlling said forward projection processor to perform forward projection operations for each respective angle of projection;

(iii) updating each image space data vector to provide an updated image space data vector for each respective angle of projection based on the corresponding image space data vector and the corresponding forward projected vector; and

(iv) iterating steps (c) (i), (ii), and (iii) until the updated image space data vectors converge to a desired extent with the corresponding forward projected vectors to reconstruct an image representative of the object.

2. Apparatus according to claim 1 further wherein the first SLM projection means includes a 1-D linear SLM array.

3. Apparatus according to claim 2 further wherein the first SLM projection means includes a first cylindrical lens for stretching an image from the 1-D linear SLM array along one axis.

4. Apparatus according to claim 1 further wherein the second SLM projection means includes a 2-D SLM array and the FPP CCD detecting array in the forward projection processor is 1-D.

5. Apparatus according to claim 4 further wherein the second SLM projection means includes a cylindrical lens for compressing an image from the 2-D SLM array along one axis.

6. Apparatus according to claim 1 further wherein the control means includes means for iteratively reconstructing an image of the object in the 2-D CCD in the back projection processor, and using forward projections of the iteratively reconstructed image and the image space data arrays to obtain error corrections applied to the 2-D CCD using the SLM projection means.

7. Apparatus according to claim 6 further wherein the control means controls the first image rotation means and second image rotation means to perform corresponding forward and backward projections for selected angles of projection.

8. Apparatus according to claim 1 further wherein the control means includes means for using state vectors representative of the image to reconstruct an image of the object in the 2-D CCD in the back projection processor, and for using forward projections of the reconstructed image and the image space data vectors to generate updated state vectors.

9. Apparatus according to claim 8 further wherein the control means controls the first image rotation means and second image rotation means to perform corresponding forward and backward projections for selected angles of projection.

10. A method for reconstruction of a two dimensional (2D) image based on a plurality of one dimensional (1-D) image projections of an object obtained from an imaging device, the 1-D image projections taken at different projection angles, the 1-D image projections represented by corresponding image space data vectors, comprising the steps of:

(a) performing back projection operations using a first spatial light modulator (SLM) device to back project a 1-D image, corresponding to each image space data vector, onto a 2-D charge coupled device (CCD) detecting array at each of said different projection angles, the back projection smearing the 1-D image on the 2-D CCD to provide back projected data representative of a 2-D image;

(b) performing forward projection operations using a second SLM device to forward project a 2-D image, corresponding to said back projected data, onto a CCD at each of said different projection angles to provide forward projected data vectors corresponding to the 1-D image space data vectors;

(c) updating each 1-D image space data vector for each respective angle of projection based on the corresponding 1-D image space data vector and the corresponding forward projected data vector; and

(d) reconstructing an image of the object by iterating steps (a), (b), and (c) until the updated image space data vectors converge to a desired extent with the corresponding forward projected data vectors.

11. A method according to claim 10 further wherein the first SLM device is a 1-D linear SLM array.

12. A method according to claim 11 further wherein the back projection operation includes the step of stretching an image obtained from the 1-D linear SLM array onto the 2-D CCD using a first cylindrical lens.

13. A method according to claim 10 further wherein the second SLM device is a 2-D SLM array.

14. A method according to claim 13 further wherein the forward projection operation uses a cylindrical lens to compress an image obtained from the 2-D SLM array onto the FP CCD.

15. A method according to claim 10 further including the steps of iteratively reconstructing an image of the object in the 2-D CCD using back projection operations, and using forward projections of the iteratively reconstructed image and the 1-D image space data vectors to obtain error corrections applied to the 2-D CCD during back projection operations.

16. A method according to claim 15 further including the steps of rotating the first SLM device relative to the 2-D CCD and rotating the second SLM device relative to the 1-D CCD to obtain forward projections and back projections at selected projection angles.

17. A method according to claim 10 further including the steps of using state vectors representative of the image to reconstruct an image of the object in the 2-D CCD, and using forward projections of the reconstructed image and the image space data vectors to generate updated state vectors.

18. A projection iterative reconstruction method for iterative reconstruction of a two-dimensional (2-D) image of an object based on a plurality of one-dimensional (1-D) image projection signals obtained from an imaging device, the projections taken at different projection angles, the image signals providing corresponding image data, said method comprising the steps of:

(a) providing a plurality of state vectors representing the image projection signals at all angles of projection;

(b) calculating a reconstructed image vector from the plurality of state vectors;

(c) forward projecting the reconstructed image vector to obtain a forward projected vector corresponding to each state vector;

(d) updating each state vector for each angle of projection based on the corresponding state vector and the corresponding forward projected vector;

(e) back projecting the updated state vectors at all angles of projection to generate back projections and summing the back projections to generate an updated reconstructed image vector;

(f) iterating steps (c), (d), and (e) until the updated state vectors converge with the corresponding forward projected vectors to a desired extent; and

(g) displaying a representation of the object based on the updated reconstructed image vector.

19. A method according to claim 18 wherein the step (d) of updating the state vectors further includes calculating the updated state vectors using a conjugate gradient technique.

20. A method according to claim 18 wherein the step (e) of back projecting and summing further includes using a spatial light modulator and a charge coupled device to update the reconstructed image vector, and wherein the step (c) of forward projecting farther includes using a spatial light modulator and a charge coupled device to obtain forward projected vectors.

21. A method according to claim 20 further wherein the spatial light modulator and charge coupled device used in the back projection of step (e) are one-dimensional and two-dimensional arrays, respectively, and the spatial light modulator and charge coupled device used in the forward projection of step (c) are two-dimensional and one-dimensional arrays, respectively.

22. A method according to claim 18 further wherein the step (d) of updating the state vectors uses a short-order FIR filter to speed up the convergence.

23. A method according to claim 18 further including the steps of using the imaging device to image the object and provide the image projection signals, and converting the image projection signals to the image data.

24. A method according to claim 18 or 23 wherein the imaging device is an x-ray imaging device having x-ray detectors, and the image projection signals are output by the x-ray detectors.

25. A method according to claim 18 further wherein the reconstructed image vector is the difference between a known portion of the image of the object and a fully reconstructed image of the object, and further including the step of adding the reconstructed image vector to the known portion to obtain the fully reconstructed image.

26. A method according to claim 18 further wherein:

in step (a) the state vectors are M×

1 vectors b_q (where M is an integer, and q denotes the projection angle of the vector);

in step (b) the reconstructed image vector is a N×

1 vector x(k) representing a reconstructed image at the kth iteration (where N is an integer and k is an integer representing the number of the iteration), and a plurality of M×

N matrices A_q represent the mapping of an image into a corresponding projection for a qth angle of projection;

in step (c) the forward projecting of a reconstructed image vector x(k-1 ) includes the step of performing a Radon transform to obtain a vector A_q x(k-1);

in step (d) a state vector t(k)_q is based on a corresponding previous state vector t(k-1 )_q and a corresponding vector A_q x(k-1), and the state vector t(k)_q represents the state of the image generated on the kth iteration;

in step (e) the back projecting of an updated state vector t(k)_q provides A_q^T t(k)_q, and the summed back projections provide an updated image x(k), where

x(k)=β

A_i^T t(k)_i summed over all projection angles i, where i=1 to q, and β

is a scalar representing a gain factor; and

in step (f) the updated reconstructed image x(k) is used as x(k-1 ) for the next iteration.

27. A method according to claim 26, further wherein t(k)_q =t(k-1)_q +α

[b_q -A_q x(k-1)], and α

is a scalar representing a gain factor.