Apparatus and method for determining measure of similarity between images

• US 20050049477A1
• Filed: 08/29/2003
• Published: 03/03/2005
• Est. Priority Date: 08/29/2003
• Status: Active Grant

• Current Assignee: Accuray Incorporated
• Original Assignee: Accuray Incorporated
• Inventors: Kuduvalli, Gopinath, Fu, Dongshan
• Granted Patent: US 7,187,792 B2
• Time in Patent Office: Days
• US Class Current: 600/407
• CPC Class Codes: A61N 2005/1061 using an x-ray imaging system having a separate imaging source
• CPC Class Codes: A61N 2005/1062 using virtual X-ray images, e.g. digitally reconstructed radiographs [DRR]
• CPC Class Codes: A61N 5/1049 for verifying the position of the patient with respect to the radiation beam
• CPC Class Codes: A61N 5/1067 in real time, i.e. during treatment
• CPC Class Codes: A61N 5/107 in real time, i.e. during treatment
• CPC Class Codes: G01N 23/04 and forming images of the material
• CPC Class Codes: G06T 2207/10124 Digitally reconstructed radiograph [DRR]
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06T 7/32 using correlation-based methods
• CPC Class Codes: G06T 7/38 Registration of image sequences

Abstract

A method and system is presented in image-guided radiosurgery for determining the measure of similarity of two digital images, for example a 2D x-ray image and a 2D DRR synthesized from 3D scan data. A two-dimensional array of pixel values of a difference image is formed by subtracting each pixel value of the second image from the corresponding pixel value of the first image. The pattern intensity function is constructed by taking the summation of asymptotic functions of the gradients of the difference image. The neighborhood R is defined so as to allow the gradients of the difference image to be considered in at least four directions.

Claims

1. A method in image-guided surgery for determining the measure of similarity of a first image of an object and a second image of said object, the method comprising:

a. forming a difference image by subtracting the corresponding pixel values of the second image from each pixel value of the first image;

wherein said first image is an x-ray image of said object generated in near real time, and said second image is a DRR (digitally reconstructed radiograph) synthesized from previously generated 3D scan data of said object; and

b. forming a pattern intensity function by summing asymptotic functions of the gradients of said difference image over all the pixels within a neighborhood R;

wherein said neighborhood R is defined so that said gradients of said difference image can be considered in at least four directions.

2. A method in accordance with claim 1, wherein said pattern intensity function is an asymptotic function of the gradients of said difference image.

3. A method in accordance with claim 1, wherein said first and second images, and said difference image, are digital images.

4. A method in accordance with claim 1, wherein said first and second images are discretized images respectively characterized by a first and a second 2D (two-dimensional) array of pixel values;

and

wherein said difference image is a discretized image characterized by a third 2D array of pixel values.

5. A method in accordance with claim 4, wherein each pixel value of an image is a number representative of the intensity of said image at a corresponding 2D array element.

6. A method in accordance with claim 4, wherein the number of rows of said first array is equal to the number of rows of said second array and said third array, and the number of columns of said first array is equal to the number of columns of said second array and said third array.

7. A method in accordance with claim 4, wherein the number of rows and columns of said first and second arrays is about 512.

8. A method in accordance with claim 1, wherein said x-ray image of said target is obtained by transmitting through said target an imaging beam having a known intensity and a known position and angle relative to said target, and wherein said 2D DRR (digitally reconstructed radiographs) representing the radiographic image of said target that would be obtained with said imaging beam at said known intensity, position and angle, if said target were positioned in accordance with said 3D scan data.

9. A method in accordance with claim 4,

wherein the pixel value for each image represents the intensity of said image, and wherein the pixel value at the i-th row and j-th column of said third array of pixel values for said difference image is given by:

I_dif(i,j)=I_Live(i,j)−

I_DRR(i,j),

wherein I_Live(i,j) represents the (i,j)-th pixel value of a real-time x-ray image of said object, and I_DRR(i,j) represents the (i,j)th pixel value of a digitally reconstructed image of said object synthesized from previously generated 3D scan data of said object.

10. A method in accordance with claim 1,

wherein said pattern intensity function is characterized by a mathematical formulation given by:

$\underset{i,j}{\overset{\text{\hspace{1em}<br><br>}}{\sum <br><br>}}<br><br>\underset{k,l⋐<br><br>R}{\overset{\text{\hspace{1em}<br><br>}}{\sum <br><br>}}<br><br>\frac{{\mathrm{\sigma <br><br>}}^2}{{\mathrm{\sigma <br><br>}}^2+{\left(I_{\mathrm{dif}}<br><br>\left(i,j\right)-I_{\mathrm{dif}}<br><br>\left(i+k,j+l\right)\right)}^2},$

where I_dif(i, j) represents the array of pixel values for said difference image, where σ

is a weighting constant for weighting said function, and where R is a neighborhood defined around the pixel (ij) as a center point.

11. A method in accordance with claim 10, wherein σ

is from about 4 to about 16.

12. A method in accordance with claim 1, wherein said at least four directions comprise:

a. a substantially horizontal direction;

b. a substantially vertical direction;

c. a diagonal direction of about 45 degrees; and

d. a diagonal direction of about −

45 degrees.

13. A method in accordance with claim 12, Wherein the result of the sampling of said pattern intensity in said neighborhood R, the pattern intensity function is given by:

$\begin{array}{c}\underset{i,j}{\overset{\text{\hspace{1em}<br><br>}}{\sum <br><br>}}<br><br>\frac{{\mathrm{\sigma <br><br>}}^2}{{\mathrm{\sigma <br><br>}}^2+{\left(I_{\mathrm{dif}}<br><br>\left(i,j\right)-I_{\mathrm{dif}}<br><br>\left(i,j-1\right)\right)}^2}+\underset{i,j}{\overset{\text{\hspace{1em}<br><br>}}{\sum <br><br>}}<br><br>\frac{{\mathrm{\sigma <br><br>}}^2}{{\mathrm{\sigma <br><br>}}^2+{\left(I_{\mathrm{dif}}<br><br>\left(i,j\right)-I_{\mathrm{dif}}<br><br>\left(i-1,j\right)\right)}^2}+\\ \underset{i,j}{\overset{\text{\hspace{1em}<br><br>}}{\sum <br><br>}}<br><br>\frac{{\mathrm{\sigma <br><br>}}^2}{{\mathrm{\sigma <br><br>}}^2+{\left(I_{\mathrm{dif}}<br><br>\left(i,j\right)-I_{\mathrm{dif}}<br><br>\left(i-1,j-1\right)\right)}^2}+\underset{i,j}{\overset{\text{\hspace{1em}<br><br>}}{\sum <br><br>}}<br><br>\frac{{\mathrm{\sigma <br><br>}}^2}{{\mathrm{\sigma <br><br>}}^2+{\left(I_{\mathrm{dif}}<br><br>\left(i,j\right)-I_{\mathrm{dif}}<br><br>\left(i-1,j+1\right)\right)}^2}\end{array}$

14. A method in accordance with claim 1, wherein said 3D scan data comprise at least one of CT scan data, MRI scan data, ultrasound scan data, and PET (positron emission tomography) data.

15. A system for determining the measure of similarity of a 2D x-ray image of an object and a 2D DRR of said object generated from previously obtained 3D scan data, said x-ray image and said DRR being discretized images characterized by a first and second 2D array of pixel values, the system comprising:

a. means for generating 3D scan data of said object;

b. an x-ray source for generating at least one imaging beam;

c. imaging means for generating a 2D radiographic image of said object in near real time, by directing said imaging beam towards and through said object from a known location and angle and at a known intensity, and detecting said imaging beam after said beam has passed through said object;

d. a controller, including;

i) software for generating a set of 2D DRR images of said object, using said 3D scan data and said known location, angle, and intensity of said imaging beam; and

ii) software for determining the measure of similarity between said 2D x-ray image and said 2D DRR, by subtracting each pixel value of said second image from a corresponding pixel value of said first image to form a difference image, by adding the asymptotic functions of the gradients of the difference image over all the pixels within a neighborhood R;

wherein said pattern intensity function is an asymptotic function of the gradients of said difference image; and

wherein said neighborhood R is defined so that said gradients of said difference image can be considered in at least four directions.

16. A system in accordance with claim 15, wherein said at least four directions comprise:

a. a substantially horizontal direction;

b. a substantially vertical direction;

c. a diagonal direction of about 45 degrees; and

d. a diagonal direction of about −

45 degrees.

17. A system in accordance with claim 15,

wherein said pattern intensity function is given by:

$\underset{i,j}{\overset{\text{\hspace{1em}<br><br>}}{\sum <br><br>}}<br><br>\underset{k,l⋐<br><br>R}{\overset{\text{\hspace{1em}<br><br>}}{\sum <br><br>}}<br><br>\frac{{\mathrm{\sigma <br><br>}}^2}{{\mathrm{\sigma <br><br>}}^2+{\left(I_{\mathrm{dif}}<br><br>\left(i,j\right)-I_{\mathrm{dif}}<br><br>\left(i+k,j+l\right)\right)}^2},$

where I_dif(i, j) represents the array of pixel values for said difference image,

where σ

is a weighting constant, and wherein said neighborhood R uses the pixel (i,j) as a center point.