Computerized tomography scanner and method of performing computerized tomography

Claim 3

The method of claim 1 wherein the algorithm to compute the discrete Fourier transform is the FFT algorithm and the algorithm to compute the inverse discrete Fourier transform is the IFFT algorithm.

Claim 4

The method of claim 1 wherein the number of samples in each Radon projection vector r i is increased by means of interpolation, and/or the number of Radon projection vectors r i is increased by means of interpolation, before the step of computing the discrete Fourier transform is computed.

Claim 5

The method of claim 1, wherein the number of samples of each vector g i 0 is increased by means of interpolation, and/or the number of vectors g i 0 is increased by means of interpolation, after the step of computing the inverse discrete Fourier transform is performed.

Claim 6

The method of claim 1 wherein the vector r i is multiplied by a weighting function in the step of (d) ii.

Claim 9

The method of claim 6 wherein steps (h) through (j) are omitted and the function (ρ) of the convolution is divided by the discrete Fourier transform of e -x .spsp.2 / σ.sbsp.o.sup..spsp.2 for all ρ such that ρ 2 ≦R 2 for some R.

Claim 10

A computerized tomography scanner, comprising: (a) means for scanning a subject by projection thereof with a source of radiation; (b) means for sensing the projected radiation with a plurality of sensors on the side of the subject opposite the source of radiation; (c) means for processing the output of the sensors to create a matrix of samples of an image, including (d) means for computing the discrete Fourier transform r i of r i using any algorithm of order N log N, where N is the length of the vector, (e) means for multiplying jth element of r i by j, for j<N/2, and multiplying the jth element of r i by N-j, for j≧N/2, where the elements of r i are numbered 0 through N-1; (f) means for computing the inverse discrete Fourier transform g i 0 of the modified r i using any algorithm of order N log N; (g) means for computing grid g l k , for l=1, . . . , Q/2 k , by merging g 2l-1 k-1 and g k-1 2l by means of rotation of coordinates, interpolation, and addition; and (h) means for creating a visual display of the matrix of samples.

Claim 11

A computerized tomography scanner, comprising: (a) means for scanning a subject by projection thereof with a source of radiation; (b) means for sensing the projected radiation with a plurality of sensors on the side of the subject opposite the source of radiation; (c) means for processing the output of the sensors to create a matrix of samples of an image, including (d) means for computing the discrete Fourier transform r i of r i using any algorithm of order N log N, where N is the length of the vector, (e) means for multiplying jth element of r i by j, for j<N/2, and multiplying the jth element of r i by N-j, for j≧N/2, where the elements of r i are numbered 0 through N-1; (f) means for computing the inverse discrete Fourier transform g i 0 of the modified r i using any algorithm of order N log N; (g) means for computing grid g l k , for l=1, . . . , Q/2 k , by merging g 2l-1 k-1 and g k-1 2l by means of rotation of coordinates, interpolation, and addition; and (h) means for creating a visual display of the matrix of samples.

Claim 12

The apparatus of claim 8, wherein the means for computing the discrete Fourier transform is the FFT algorithm and the means for computing the inverse discrete Fourier transform is the IFFT algorithm.

Claim 13

The apparatus of claim 8, wherein the number of samples in each Radon projection vector r i is increased by means of interpolation, and/or the number of Radon projection vectors r i is increased by means of interpolation, before computing the discrete Fourier transform.

Claim 14

The apparatus of claim 8, wherein the number of samples of each vector g i 0 is increased by means of interpolation, and/or the number of vectors g i 0 is increased by means of interpolation, after computing the inverse discrete Fourier transform.

Claim 15

The apparatus of claim 8, wherein the vector r i is multiplied by a weighting function in means for multiplying.

Claim 16

A computerized tomography scanner, comprising: (a) means for scanning a subject by projection of radiation; (b) means for sensing the projected radiation with a plurality of sensors; (c) means for processing the output data of the sensors to create a matrix of samples of an image, including; (d) means for computing the convolution and backprojection of the output data; (e) means for computing a selection of the point spread functions produced by the Radon transform-convolution-backprojection suite; (f) means for computing the width σ o of the Gaussian that most closely fits the computed point spread functions according to a mathematical criterion; (g) means for computing G, the 2-D discrete Fourier transform (2DDFT) of the N×N matrix of samples of the Gaussian of width σ o ; (h) means for computing the 2-D discrete Fourier transform of the N×N image matrix using any algorithm of order N 2 logN; (i) means for dividing each component of the resulting matrix by the corresponding sample of G, with treatment of components at which G is near zero; (j) means for computing the inverse 2-D discrete Fourier transform of the resulting matrix; and (k) means for converting the matrix data to a visual display.

Claim 17

A computerized tomography scanner, comprising: (a) means for scanning a subject by projection of radiation; (b) means for sensing the projected radiation with a plurality of sensors; (c) means for processing the output data of the sensors to create a matrix of samples of an image, including; (d) means for computing the convolution and backprojection of the output data; (e) means for computing a selection of the point spread functions produced by the Radon transform-convolution-backprojection suite; (f) means for computing the width σ o of the Gaussian that most closely fits the computed point spread functions according to a mathematical criterion; (g) means for computing G, the 2-D discrete Fourier transform (2DDFT) of the N×N matrix of samples of the Gaussian of width σ o ; (h) means for computing the 2-D discrete Fourier transform of the N×N image matrix using any algorithm of order N 2 logN; (i) means for dividing each component of the resulting matrix by the corresponding sample of G, with treatment of components at which G is near zero; (j) means for computing the inverse 2-D discrete Fourier transform of the resulting matrix; and (k) means for converting the matrix data to a visual display.

Claim 18

A computer readable memory medium encoded with data representing a computer program for use with a computerized tomography scanner and a computer to generate a visual image by: (a) filtering samples of Radon projection sample vectors r i , where i=1, . . . , Q, by: i. computing the discrete Fourier transform r i of r i using an algorithm of order N log N, where N is the length of the vector; ii. multiplying the jth element of r i by j, for j<N/2, and multiplying the jth element of r i by N-j, for j≧N/2, where the elements of r i are numbered 0 through N-1; and iii. computing the inverse discrete Fourier transform g i 0 of the modified r i using any algorithm of order N log N; and (b) performing the kth level of merges, for k=1, . . . , log 2 Q, by computing the grid g l k , for l=1, . . . , Q/2 k , by merging g 2l-1 k-1 and g k-1 2l by means of rotation of coordinates, interpolation, and addition.

Claim 19

A computer readable memory medium encoded with data representing a computer program for use with a computerized tomography scanner and a computer to generate a visual image by: (a) filtering samples of Radon projection sample vectors r i , where i=1, . . . , Q, by: i. computing the discrete Fourier transform r i of r i using an algorithm of order N log N, where N is the length of the vector; ii. multiplying the jth element of r i by j, for j<N/2, and multiplying the jth element of r i by N-j, for j≧N/2, where the elements of r i are numbered 0 through N-1; and iii. computing the inverse discrete Fourier transform g i 0 of the modified r i using any algorithm of order N log N; and (b) performing the kth level of merges, for k=1, . . . , log 2 Q, by computing the grid g l k , for l=1, . . . , Q/2 k , by merging g 2l-1 k-1 and g k-1 2l by means of rotation of coordinates, interpolation, and addition.

Claim 20

The computer readable memory medium of claim 14 in the form of a floppy disk, an ASIC, computer chip or circuit.

Claim 21

The computer readable memory medium according to claim 14 wherein the number of samples in each Radon projection vector r i is increased by means of interpolation, and/or the number of Radon projection vectors r i is increased by means of interpolation, before the step of computing discrete Fourier transform is computed.

Claim 22

The computer readable memory medium according to claim 14 wherein the number of samples of each vector g i 0 is increased by means of interpolation, and/or the number of vectors g i 0 is increased by means of interpolation, after the step of comprising inverse discrete Fourier transform is performed.

Claim 23

The computer readable memory medium according to claim 14 wherein the vector r i is multiplied by a weighting function in the step of multiplying discrete Fourier transform is performed.

Claim 24

The computer readable memory medium encoded with data representing a computer program that can cause a computer to function to execute the method of: (a) computing a selection of the point spread functions produced by the Radon transform-convolution-backprojection suite; p1 (b) computing the width σ o of the Gaussian that best fits the computed point spread functions according to a mathematical criterion; (c) computing G, the 2-D discrete Fourier transform (2DDFT) of the N×N matrix of samples of the Gaussian of width σ o ; (d) computing the 2DDFT of the N×N image matrix using an algorithm of order N 2 logN; (e) dividing each component of the resulting matrix by the corresponding sample of G, with special treatment of components at which G is near zero; and (f) computing the inverse 2-D discrete Fourier transform of the resulting matrix.

Claim 25

The computer readable memory medium encoded with data representing a computer program that can cause a computer to function to execute the method of: (a) computing a selection of the point spread functions produced by the Radon transform-convolution-backprojection suite; p1 (b) computing the width σ o of the Gaussian that best fits the computed point spread functions according to a mathematical criterion; (c) computing G, the 2-D discrete Fourier transform (2DDFT) of the N×N matrix of samples of the Gaussian of width σ o ; (d) computing the 2DDFT of the N×N image matrix using an algorithm of order N 2 logN; (e) dividing each component of the resulting matrix by the corresponding sample of G, with special treatment of components at which G is near zero; and (f) computing the inverse 2-D discrete Fourier transform of the resulting matrix.

Claim 26

The computer readable memory medium of claim 19 in the form of a floppy disk, an ASIC, computer chip or circuit.