Cone beam scanner using oblique surface reconstructions

• US 6,343,108 B1
• Filed: 06/16/2000
• Issued: 01/29/2002
• Est. Priority Date: 06/18/1999
• Status: Expired due to Fees

• Current Assignee: Marconi Medical Systems, Inc. (Koninklijke Philips N.V.)
• Original Assignee: Philips Medical Systems (Cleveland) Incorporated (Koninklijke Philips N.V.)
• Inventors: Heuscher, Dominic J.
• Primary Examiner(s): BRUCE, DAVID VERNON
• Application Number:
  US09/595,858
• Time in Patent Office: 592 Days
• Field of Search: 378/4, 378/15, 378/901
• US Class Current: 378/4
• CPC Class Codes: A61B 6/027 characterised by the use of a particular data acquisition trajectory, e.g. helical or spiral
• 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: Y10S 378/901 Computer tomography program or processor

Abstract

Data collected from a cone beam scanner is reconstructed by defining a plurality of oblique surfaces (44) which are reconstructed into a reconstruction cylinder (C). An interpolator (52) identifies non-redundant rays that intersect a center point of each oblique surface (70) along with rays tangent to surface rings on each oblique surface (72, 74). Data from the non-redundant rays is weighted by a first data processor (54). A second data processor (56) convolves the weighted data and passes it to a backprojector (58) which backprojects the data into an image memory (60). The oblique surface reconstruction technique facilitates the use of conventional two-dimensional convolution and backprojection techniques that enjoy relative computational simplicity and efficiency as well as three-dimensional reconstruction techniques that use a minimum number of projections. In addition, the oblique surface reconstruction technique facilitates accurate reconstruction of cone beam projections with over ten times the area of current multi-ring scanners.

Claims

1. A method of volumetric image reconstruction from partial cone beam data comprising:

(a) collecting the partial cone beam data in two-dimensional arrays at a plurality of sampling positions, said collected data corresponding to rays of radiation diverging in two dimensions from a common vertex as the vertex travels along a continuous path;

(b) defining a plurality of oblique surfaces throughout a region of interest;

(c) identifying rays of radiation which pass through the plurality of oblique surfaces;

(d) reconstructing data from the identified rays into a reconstruction cylinder having an axis along a z-direction; and

(e) generating a volume data set from the reconstructed oblique surface data.

2. The method according to claim 1 further comprising:

rebinning the collected data into a wedge beam format, wherein sets of parallel rays are grouped to define planes of radiation that angularly diverge from a common axis.

3. The method according to claim 1, wherein step (c) includes:

at each sampling position, identifying non-redundant rays of radiation traversing each oblique surface throughout the region of interest.

4. The method according to claim 1, wherein step (c) includes:

at each sampling position, identifying rays of penetrating radiation that intersect at a center point of each oblique surface.

5. The method according to claim 4, wherein step (c) further includes:

identifying pixels on the oblique surface at least one pixel out from the center point, said pixels at least one pixel out from the center point forming a first surface ring; and

identifying rays of penetrating radiation which are tangent to the first surface ring.

6. The method according to claim 5, wherein the step (c) further includes:

iteratively identifying pixels outward from the center point of each oblique surface to a reconstruction cylinder radius, said pixels forming a plurality of oblique surface rings; and

iteratively identifying rays of penetrating radiation tangent to the surface rings.

7. The method according to claim 6, wherein step (c) includes interpolating closest adjacent rays between the center point and the reconstruction cylinder radius where partial cone beam data is not collected.

8. The method according to claim 7, wherein step (c) further includes weighting the interpolated rays by a weighting function such that the weights of opposing rays sum to unity.

9. The method according to claim 8, wherein weighting the interpolated rays includes:

applying plus-region weighting along the z-direction; and

applying cone beam angle weighting.

10. The method according to claim 9, wherein the vertex travels along a helical path.

11. The method according to claim 10, wherein step (d) includes:

computing a two-dimensional convolution of the weighted data; and

two-dimensionally backprojecting the weighted convolution into a volumetric image memory.

12. The method according to claim 1, said method further comprising:

defining a plurality of path segments along the path of the vertex, said path segments being at least 180°

plus a fan angle of the cone beam; and

,

in step (b), each of the defined oblique surfaces is derived from the corresponding path segment.

13. The method according to claim 12, wherein the vertex travels along a helical path.

14. The method according to claim 7, wherein the spacing of the reconstructed surfaces along the longitudinal axis of the reconstruction cylinder is not greater than one-half of an effective longitudinal axis resolution.

15. The method according to claim 1, wherein step (e) includes:

interpolating matrices corresponding to the reconstructed oblique surface data along the longitudinal direction to form the volume data set; and

filtering the resulting volume data set along the longitudinal direction in order to define an effective longitudinal axis resolution.

16. A method of radiographic diagnostic imaging comprising:

(a) generating penetrating radiation;

(b) receiving the penetrating radiation with two-dimensional radiation detectors along a plurality of divergent rays, the rays being focused at a common origin vertex and diverging in two dimensions;

(c) rotating the vertex along at least an arc segment of a helical path;

(d) sampling the radiation detectors at a plurality of angular increments along the helical arc segment to generate a plurality of two-dimensional projection views, each view including a two-dimensional array of data values, each data value corresponding to one of the divergent rays;

(e) defining a first and a last oblique surface formed by the intersection of a cone beam of penetrating radiation and a region of interest;

(f) defining a plurality of additional oblique surfaces, said plurality of oblique surfaces being at least one of rotated and translated with respect to the first oblique surface;

(g) weighting projection views corresponding to each oblique surface;

(h) computing a two-dimensional convolution of the projection view data; and

(i) two-dimensionally backprojecting the convolved projection data corresponding to each oblique surface into a volumetric image memory.

17. The method according to claim 16, wherein steps (e) and (f) include:

identifying the rays of penetrating radiation that intersect at a center point of the oblique surface;

iteratively identifying pixels outward from the center point to an outward radius of a reconstruction cylinder, said outward pixels forming a plurality of oblique surface rings; and

iteratively identifying rays of penetrating radiation which are tangent to each of the plurality of oblique surface rings.

18. The method according to claim 17, wherein steps (e) and (f) further include:

interpolating closest adjacent rays tangent to the oblique surface rings between the center point and the outer radius of the reconstruction cylinder where penetrating radiation was not detected.

19. A method of selecting non-redundant rays of penetrating radiation during a computed tomography scan, said non-redundant rays forming a plurality of oblique surfaces for two-dimensional reconstruction into a volumetric image representation, said method comprising:

at each angular orientation about an examination region, selecting detected rays of penetrating radiation that intersect a geometric center point of each oblique surface; and

iteratively identifying rays of penetrating radiation which are tangent to surface rings extending outward from the geometric center point of each oblique surface.

20. The method according to claim 19, said method further comprising:

where penetrating radiation data does not exists, interpolating closest adjacent rays tangent to the oblique surface rings between the center point and an outer radius of region of interest.

21. The method according to claim 20, wherein each oblique surface is at least one of translated and rotated with respect to a first oblique surface.

22. A computed tomography (CT) scanner comprising:

a first gantry which defines an examination region;

a rotating gantry mounted on the first gantry for rotation about the examination region;

a source of penetrating radiation arranged on the rotating gantry for rotation therewith, said source of penetrating radiation emitting a cone-shaped beam of radiation that passes through the examination region as the rotating gantry rotates;

a subject support which holds a subject being examined at least partially within the examination region, wherein at least one of the first gantry and the subject support is translated such that the subject passes through the examination region while the rotating gantry is rotated and the source of penetrating radiation follows a helical path relative to the subject;

a two-dimensional array of radiation detectors arranged to receive the radiation emitted from the source of penetrating radiation after it has traversed the examination region;

a reconstruction processor which reconstructs images of the subject from data collected by the two-dimensional array of radiation detectors, said reconstruction processor including;

a control processor which defines a plurality of oblique surfaces, said oblique surfaces being defined by the intersection of the cone-shaped beam of radiation and a portion of the subject;

an interpolator which identifies non-redundant rays of penetrating radiation that pass through the oblique surfaces;

a first data processor which weights the data corresponding to the identified non-redundant rays;

a second data processor which receives the data from the first data processor and performs a two-dimensional convolution on the data; and

,

a backprojector which receives the data from the second data processor and two-dimensionally backprojects it into an image memory; and

,

a human-viewable display which accesses the image memory to display reconstructed images of the subject.

23. The CT scanner according to claim 22, wherein the first data processor applies (i) plus region weighting, and (ii) cone beam angle weighting to the data.

24. The CT scanner according to claim 22, wherein the interpolator:

at each sampling position, identifies rays of penetration radiation that intersect at a center point of each oblique surface;

iteratively identifies pixels outward from the center point of each oblique surface to an outward radius of a reconstruction cylinder, said outward pixels forming a plurality of oblique surface rings; and

,

iteratively identifies rays of penetrating radiation which are tangent to each of the plurality of oblique surface rings.

25. The CT scanner according to claim 23, wherein the interpolator interpolates nearest adjacent rays which are tangent to the oblique surface rings.