Post-processing technique for reducing metallic clip artifacts in CT images

• US 5,243,664 A
• Filed: 09/16/1991
• Issued: 09/07/1993
• Est. Priority Date: 09/16/1991
• Status: Expired due to Term

• Current Assignee: Picker International, Inc. (Marconi Holdings Company Limited)
• Original Assignee: Picker International, Inc. (Marconi Holdings Company Limited)
• Inventors: Tuy, Heang K.
• Primary Examiner(s): Mancuso, Joseph
• Application Number:
  US07/760,402
• Time in Patent Office: 722 Days
• Field of Search: 382/6, 382/54, 382/57, 364/413.13, 364/413.14, 364/413.16, 364/413.18, 364/413.19
• US Class Current: 382/130
• CPC Class Codes: G06T 11/005 Specific pre-processing for tomographic reconstruction, e.g. calibration, source positioning, rebinning, scatter correction, retrospective gating

Abstract

Radiation from an x-ray source (14) is collected by radiation detectors (16) to generate an original set of CT line integral data (18). The original CT data is subject to a first order correction and reconstructed (20) to generate a first order corrected image representation (22). A first artifact image representation (26) is generated (24) from the first order corrected image. The artifact image representation and the first order corrected image representation are subtractively combined (28) to generate a second order corrected image representation (30). A bone and high density clip image representation (142) is generated from the second order corrected image. The bone and clip image representation is forward projected (144) to generate a first high contrast line integral data set (146). High contrast model line integral data corresponding to rays which traverse the high density clip are topologically transposed (148) for corresponding rays of the original CT data. The topologically transposed CT data is reconstructed (160) to generate a third order corrected image representation (34). A second artifact image representation (38) is derived (36) from the third order image representation. The second artifact image representation and the third order corrected image representation are subtractively combined (40) to produce a fourth order corrected image representation (42) which is displayed on a video monitor (44) or the like.

Claims

1. A method of diagnostic imaging in which streak artifacts from high density materials are reduced, the method comprising:

passing penetrating radiation through a subject along a plurality of rays;

converting an intensity of radiation which has traversed each ray through the subject into CT data values, each data value being indicative of a line integral of radiation absorptive or transmissive properties of material along the corresponding ray;

reconstructing a first image representation from the CT data values;

deriving a first artifact image representation from the CT data values, which first artifact image representation is indicative of artifacts attributable to the high density material in the subject;

combining the first image representation and the first artifact image representation to generate a second order corrected image representation indicative of the slice through the subject with a correction made for the high density material artifact.

2. The method as set forth in claim 1 further including:

generating a high contrast model from the second order corrected image representation, which high contrast model represents an image of bone and high density materials with a uniform water background;

generating a set of high contrast model line integral data corresponding to the high contrast model;

topologically transposing high contrast model line integral data corresponding to rays which intersect the high density material for CT data values of corresponding rays of the CT data to generate a topologically translated data set;

reconstructing the topologically translated data set into a third order corrected image representation.

3. A method in which streak artifacts from high density materials are reduced, the method comprising:

passing penetrating radiation through a subject along a plurality of rays;

converting an intensity of radiation which has traversed each ray through the subject into CT data values, each data value being indicative of a line integral of radiation absorptive or transmissive properties of material along the corresponding ray;

reconstructing a first image representation from the CT data values;

deriving a first artifact image representation from the CT data values, which first artifact image representation is indicative of artifacts attributable to the high density material in the subject;

combining the first image representation and the first artifact image representation to generate a second order corrected image representation indicative of the slice through the subject with a correction made for the high density material artifact;

generating a high contrast model from the second order corrected image representation, which high contrast model represents an image of bone and high density materials against a uniform water background;

generating a set of high contrast model line integral data corresponding to the high contrast model;

topologically transposing high contrast model line integral data corresponding to rays which intersect the high density material for CT data values of corresponding rays of the CT data to generate a topologically translated data set;

reconstructing the topologically translated data set into a third order corrected image representation;

generating a second artifact image representation from the third order corrected image representation and the high contrast model line integral data;

subtractively combining the second artifact image representation with the third order corrected image representation to generate a fourth order corrected image representation.

4. The method as set forth in claim 3 further including connecting the fourth order corrected image representation into a man-readable display.

5. The method as set forth in claim 4 wherein the step of generating the first image representation includes:

reconstructing an image representation from the CT data values;

thresholding the reconstructed image representation to generate an image representation of only the high density material;

forward projecting the high density material image to generate a set of high density line integral data corresponding to the high density material image representation;

subtractively combining the high density material line integral data with the CT data values to generate a revised data set;

compensating for beam hardness effects by adding a deficiency compensation derived from the first revised data set to the CT data values;

reconstructing the deficiency compensated CT data values to generate the first image representation.

6. The method as set forth in claim 5 wherein the deficiency compensating step includes:

determining a linear interpolation of line integral data corresponding to rays which intersect opposite extremes of the high density material to generate a linear interpolated value corresponding to each ray which intersects the high density material;

determining a differential value for the line integral data corresponding to each ray which intersects the high density material;

for each ray which intersects the high density material, subtracting the linear interpolation value and the differential value;

comparing the differential and linear interpolation values of the corresponding ray;

in response to the linear interpolation value exceeding the differential value, adding the difference to the corresponding ray of the original CT data.

7. The method as set forth in claim 4 wherein the step of generating the first artifact image representation includes:

thresholding the first order corrected image representation to generate a high contrast model image representing an image of bone and high density material;

forward projecting the high contrast model to generate a set of line integral data corresponding to the high contrast model;

determining a linear interpolation of line integral data corresponding to rays which intersect opposite extremes of the high density material to generate a linear interpolated value corresponding to each ray which intersects the high density material;

determining a differential value for the line integral data corresponding to each ray which intersects the high density material;

for each ray which intersects the high density material, subtracting the corresponding linear interpolation value and the differential value;

comparing the differential and linear interpolation values corresponding to each ray;

in response to the linear interpolation value exceeding the differential value, adding the difference to the high contrast model line integral corresponding ray;

reconstructing the data of the high contrast model into a corrected high contrast model image representation;

subtractively combining the corrected high contrast model image representation and the high contrast model to generate the first artifact image representation.

8. The method as set forth in claim 4 wherein the topological transformation step includes:

discarding the data values of the CT data set corresponding to rays passing through the high density material;

adjusting an amplitude of line high contrast model integral data in accordance with values of the CT data values corresponding to rays immediately adjacent opposite sides of the high density material;

inserting the amplitude adjusted high contrast model line integral data for the corresponding rays of the CT data values.

9. The method as set forth in claim 4 wherein the step of generating the second artifact image representation includes:

thresholding the third order corrected image representation to generate a second high contrast model;

forward projecting the second high contrast model into second high contrast model line integral data set;

topologically transposing line integral data corresponding of the first high contrast model for corresponding rays of the second high contrast model line integral data;

reconstructing a corrected high contrast model image representation from the topologically transposed second high contrast model line integral data;

subtractively combining the second high contrast model image representation and the corrected high contrast model image representation to generate the second artifact image.

10. A method of diagnostic imaging in which streak artifacts from a high density material are reduced:

passing penetrating radiation through a subject along a plurality of rays;

converting an intensity of radiation which has traversed each ray through the subject into CT data values, each data value being indicative of a line integral of radiation absorptive or transmissive properties of material along the corresponding ray;

reconstructing a first image representation from the CT data values;

generating a high contrast model from the first image representation;

forward projecting the high contrast model to generate high contrast model line integral data;

replacing portions of the high contrast model line integral data that corresponds to rays passing through the high density material with corresponding CT data values;

reconstructing the high contrast model line integral data with replaced portions into a second image representation;

subtractively combining the second image representation and the high contrast model to create an artifact image representation;

combining the first image representation and the first artifact image representation to generate a second order corrected image representation indicative of the slice through the subject with a correction made for the high density material artifact.

11. The method as set forth in claim 10 further including before the forward projecting step, replacing portions of the high contrast model that correspond to bone and clip with corresponding portions of the first image representation.

12. A method of diagnostic imaging in which streak artifacts from a high density clip are reduced, the method comprising:

converting an intensity of penetrating radiation which has traversed each of a plurality of rays through a subject into subject data indicative of sums of radiation absorptive or transmissive properties of material along the corresponding rays;

reconstructing a streak artifact contaminated image representation from the subject data;

deriving an artifact image representation from the subject data, which artifact image representation is indicative of artifacts attributable to the high density clip;

subtractively combining the streak artifact contaminated image representation and the artifact image representation to generate a streak artifact corrected image representation.

13. The method as set forth in claim 12 wherein the artifact image representation deriving step includes:

generating a high contrast model from the streak artifact contaminated image representation;

forward projecting the high contrast model to generate high contrast model line integral data;

replacing portions of the high contrast model line integral data that corresponds to rays passing through the clip with corresponding subject data;

reconstructing the high contrast model line integral data with replaced portions into a revised high contrast model image representation;

subtractively combining the revised high contrast model image representation and the high contrast model to create the artifact image representation.

14. The method as set forth in claim 12 wherein the step of generating the artifact image representation includes:

thresholding the streak artifact contaminated representation to generate a high contrast model;

forward projecting the high contrast model into a high contrast model line integral data set;

topologically transposing substitute line integral data for a portion of the high contrast model line integral data which corresponds to rays that intersect the clip;

reconstructing a corrected high contrast model image representation from the topologically transposed high contrast model line integral data;

subtractively combining the high contrast model and the corrected high contrast model image representation to generate the artifact image representation.

15. The method as set forth in claim 12 further including the step of displaying the streak artifact corrected image representation.

16. A method of diagnostic imaging in which streak artifacts from a high density clip are reduced, the method comprising:

passing penetrating radiation through a subject along a plurality of rays;

converting an intensity of radiation which has traversed each ray through the subject into line integral data indicative of line integrals of radiation absorptive or transmissive properties of the subject along the corresponding rays;

reconstructing an image representation from the subject line integral data;

generating a high contrast model from the image representation, which high contrast model represents bone and the clip;

forward projecting a set of high contrast model line integral data from the high contrast model;

topologically transposing high contrast model line integral data corresponding to rays which intersect the clip for the subject line integral data of corresponding rays;

reconstructing the topologically translated subject line integral data into a corrected image representation.

17. The method as set forth in claim 16 wherein the topological transformation step includes:

discarding portions of the subject line integral data corresponding to rays passing through the clip;

adjusting the amplitude of the high contrast model line integral data corresponding to rays through the clip in accordance with subject line integral;

inserting the amplitude adjusted high contrast model line integral data for the discarded portions of the subject line integral data.

18. A method of diagnostic imaging in which streak artifacts from a high density clip are reduced, the method comprising:

converting an intensity of penetrating radiation which has traversed each of a plurality of rays through a subject into line integral data indicative of line integrals of radiation absorptive or transmissive properties of the subject along the corresponding rays;

reconstructing an image representation from the subject line integral data;

thresholding the reconstructed image representation to generate an image representation of at least the clip;

forward projecting the clip image representation to generate clip line integral data;

subtractively combining the clip and subject line integral data to generate revised line integral data;

compensating for beam hardness effects by adding a deficiency compensation derived from the revised line integral data to the subject line integral data;

reconstructing the deficiency compensated subject line integral data to generate a corrected subject image representation.

19. The method as set forth in claim 18 wherein the thresholding step includes:

thresholding the reconstructed image representation to generate an image representation of the clip and bone, whereby the clip line integral data is based on both the bone and the clip.

20. A method of diagnostic imaging comprising:

passing penetrating radiation along a plurality of rays through a portion of a subject that includes a high density clip;

converting an intensity of radiation which has traversed each ray into line integral data indicative of line integrals of radiation absorptive or transmissive properties of the subject along the corresponding rays;

reconstructing an image representation from the subject line integral data;

thresholding the reconstructed image representation to generate an image representation of at least the clip;

forward projecting the clip image representation to generate clip line integral data;

subtractively combining the clip and subject line integral data to generate revised line integral data;

determining a linear interpolation of portions of the revised line integral data corresponding to rays which intersect opposite extremes of the clip to generate a linear interpolated data value corresponding to each ray of the revised line integral data intersects the clip;

determining a differential value from the revised line integral data corresponding to each ray that intersects the clip;

for each ray which intersects the clip, subtracting the corresponding linear interpolated value and the differential value;

in response to the linear interpolation value exceeding the differential value of the corresponding ray, adding the difference to the subject line integral value to generate deficiency compensated subject line integral data;

reconstructing the deficiency compensated subject line integral data to generate a corrected subject image representation.

21. A method of diagnostic imaging in which streak artifacts from a high density clip are reduced, the method comprising:

passing penetrating radiation through a subject along a plurality of rays;

converting an intensity of radiation which has traversed each ray through the subject into line integral data indicative of line integrals of radiation absorptive or transmissive properties of material along the corresponding rays;

generating high contrast line integral data from the subject line integral data, which high contrast line integral data corresponds to higher density portions of the subject and the clip;

determining a linear interpolation of a portion of the high contrast line integral data corresponding to rays which intersect opposite extremes of the clip to generate a linear interpolated data value corresponding to each ray that intersects the clip;

determining a differential value from the high contrast line integral data corresponding to each ray that intersects the clip;

for each ray which intersects the high density material, subtractively combining the corresponding linear interpolated and differential values;

adding the difference between the linear interpolated and differential values to the high contrast model line integral of the corresponding ray, at least when the linear interpolated value exceeds the differential value;

correcting a subject image representation generated from the subject line integral data in accordance with the high contrast line integral data with the added difference.

22. The method as set forth in claim 21 wherein the correcting step includes:

generating a corrected high contrast image representation from the high contrast line integral data with the added difference;

subtractively combining the corrected high contrast model image representation with an uncorrected high contrast model image representation to generate an artifact image representation;

subtractively combining the subject image representation and the artifact image representation.

23. An apparatus for diagnostic imaging in which streak artifacts from high density materials are reduced, the apparatus comprising:

means for passing penetrating radiation through a subject along a plurality of rays;

means for converting an intensity of radiation which has traversed each ray through the subject into line integral data values, each line data value being indicative of a line integral of radiation absorptive or transmissive properties of material along the corresponding rays;

means for reconstructing a streak artifact contaminated image representation from the line integral data;

means for deriving an artifact image representation from the line integral data values, which artifact image representation is indicative of artifacts attributable to the high density material in the subject;

means for combining the streak artifact contaminated image representation with the artifact image representation to generate a corrected image representation.

24. A diagnostic imaging apparatus for generating images in which streak artifacts from high density material are reduced, the apparatus comprising:

a means for reconstructing a first image representation from line integral data values, which line integral data value are each indicative of a line integral of radiation absorptive or transmissive properties of material along one of a plurality of corresponding rays through a portion of the subject that contains the high density material;

a means for deriving an artifact image representation from the line integral data values, which artifact image representation is indicative of artifacts attributable to the high density material in the subject;

a means for combining the first image representation and the artifact image representation to generate a second order corrected image representation indicative of the slice through the subject with a correction made for the high density material artifact.

25. A diagnostic imaging apparatus for generating diagnostic images in which streak artifacts from a high density clip are reduced, the apparatus comprising:

a means or reconstructing a first image representation from line integral data values, which line integral data values are each indicative of a line integral of radiation absorptive or transmissive properties of material along one of a plurality of corresponding rays through a portion of the subject that contains the high density clip;

a means for generating a high contrast model from the first image representation, which high contrast model represents bone and the clip;

a means for forward projecting a set of high contrast model line integral data from the high contrast model;

a means for topologically transposing high contrast model line integral data corresponding to rays which intersect the clip for the subject line integral data of corresponding rays;

a means for reconstructing the topologically translated subject line integral data into a corrected image representation.

26. A diagnostic imaging apparatus for generating diagnostic image representations in which streak artifacts from a high density clip are reduced, the method comprising:

a means for passing penetrating radiation along a plurality of rays through a subject;

a means for converting an intensity of radiation which has traversed each ray into line integral data indicative of line integrals of radiation absorptive or transmissive properties of the subject along the corresponding rays;

a means for reconstructing an image representation from the subject line integral data;

a means for thresholding the reconstructed image representation to generate an image representation of at least the clip;

a means for forward projecting the clip image representation to generate clip line integral data;

a means for subtractively combining the clip and subject line integral data to generate revised line integral data;

a means for compensating for beam hardness effects by adding a deficiency compensation derived form the revised line integral data to the subject line integral data;

a means for reconstructing the deficiency compensated subject line integral data to generate a corrected subject image representation.

27. A CT scanner comprising:

a means for passing penetrating radiation through a subject along a plurality of rays;

a means for converting an intensity of radiation which has traversed each ray through the subject into line integral data indicative of line integrals of radiation absorptive or transmissive properties of material along the corresponding rays;

a means for generating high contrast line integral data from the subject line integral data, which high contrast line integral data corresponds to higher density portions of the subject and the clip;

a means for determining a linear interpolation of a portion of the high contrast line integral data corresponding to rays which intersect opposite extremes of the clip to generate a linear interpolated data value corresponding to each ray that intersects the clip;

a means for determining a differential value from the high contrast line integral data corresponding to each ray that intersects the clip;

a means for subtractively combining the corresponding linear interpolated and differential values for each ray which intersects the high density material;

a means for adding the difference between the linear interpolated and differential values to the high contrast model line integral of the corresponding ray, at least when the linear interpolated value exceeds the differential value;

a means for correcting a subject image representation generated from the subject line integral data in accordance with the high contrast line integral data with the added difference.