GPU-based image manipulation method for registration applications

• US 20050271302A1
• Filed: 04/19/2005
• Published: 12/08/2005
• Est. Priority Date: 04/21/2004
• Status: Active Grant

• Current Assignee: Siemens Corp. (Siemens AG)
• Original Assignee: Siemens Corp. (Siemens AG)
• Inventors: Geiger, Bernhard, Khamene, Ali, Vogt, Sebastian, Chefd'hotel, Christophe, Guhring, Jens
• Granted Patent: US 7,706,633 B2
• Time in Patent Office: Days
• US Class Current: 382/294
• CPC Class Codes: G06T 1/20 Processor architectures; Processor configuration, e.g. pipelining
• CPC Class Codes: G06T 15/005 General purpose rendering architectures
• CPC Class Codes: G06T 2200/28 involving image processing hardware
• CPC Class Codes: G06T 2207/10072 Tomographic images
• CPC Class Codes: G06T 2207/10116 X-ray image
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/30004 Biomedical image processing
• CPC Class Codes: G06T 7/35 using statistical methods
• CPC Class Codes: G06V 10/75 Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
• CPC Class Codes: G06V 10/955 using specific electronic processors

Abstract

Exemplary systems and methods for performing registration applications are provided. An exemplary system includes a central processing unit (CPU) for transferring a plurality of images to a graphics processing unit (GPU); wherein the GPU performs a registration application on the plurality of images to produce a registration result, and wherein the GPU returns the registration result to the CPU. An exemplary method includes the steps of transferring a plurality of images from a central processing unit (CPU) to a graphics processing unit (GPU); performing a registration application on the plurality of images using the GPU; transferring the result of the step of performing from the GPU to CPU.

Claims

1. A system for performing registration applications, comprising:

a central processing unit (CPU) for transferring a plurality of images to a graphics processing unit (GPU);

wherein the GPU performs a registration application on the plurality of images to produce a registration result, and wherein the GPU returns the registration result to the CPU.

2. The system of claim 1, wherein the registration application comprises at least one of:

an image warping application;

a local histogram computation application;

an intensity correction application; and

a similarity measure estimation application.

4. The system of claim 1, wherein the GPU comprises:

a vertex shading unit for computing texture coordinates for each vertex in a geometric shape in the plurality of images, and for interpolating the texture coordinates for each pixel in the geometric shape.

5. The system of claim 1, wherein the GPU comprises:

a negative number handling unit for handling negative numbers in the GPU.

6. The system of claim 1, wherein the GPU comprises:

a gradient value handling unit for handling gradient values in GPU.

7. The system of claim 1, wherein the GPU comprises:

a mipmap unit for averaging contents of the plurality of images.

8. The system of claim 1, wherein the GPU comprises:

a pixel buffer (Pbuffer) comprising an off-screen render buffer.

3. The system of claim 3, wherein the similarity measure estimation application comprise one of:

a sum of squared intensity differences (SSD) application;

a sum of absolute intensity differences (SAD) application;

a normalized correlation coefficient (NCC) application;

a ratio image uniformity (RIU) application;

a pattern intensity (PI) application;

a gradient conversion (GC) application;

a gradient difference (GD) application;

a sum of local normalized correlation (SLNC) application;

a variance-weighted sum of local normalized correlation (VWC) application; and

a mutual information (MC) application.

9. A method for registering images, comprising:

transferring a plurality of images from a central processing unit (CPU) to a graphics processing unit (GPU);

performing a registration application on the plurality of images using the GPU; and

transferring the result of the step of performing from the GPU'"'"'to CPU.

10. The method of claim 9, wherein the step of performing a registration application using the GPU comprises at least one of:

performing an image warping application;

performing a local histogram computation application;

performing an intensity correction application; and

performing a similarity measure estimation application.

11. The method of claim 9, wherein the step of performing a registration application on the plurality of images using GPU, comprises:

reducing the size of the plurality of images; and

performing the registration application on the size-decreased plurality of images.

12. The method of claim 11, wherein the step of reducing the size of the plurality of images, comprises:

generating a plurality of lower-resolution mipmaps based on the plurality of images.

13. The method of claim 12, wherein the step of transferring the result of the step of performing from the GPU to CPU, comprises:

transferring the lower-resolution mipmaps from the GPU to the CPU; and

performing averaging operations on the lower-resolution mipmaps.

14. The method of claim 9, wherein the step of performing a registration application using GPU, comprises:

mapping a number of similarity measures from the CPU to the GPU; and

performing one of the number of similarity measures on the plurality of images using the GPU.

15. The method of claim 9, wherein the step of performing one of the number of similarity measures, comprises one of:

performing sum of absolute intensity differences (SAD);

performing sum of squared intensity differences (SSD);

performing normalized cross correlation (NCC);

performing ratio image uniformity (RIU);

performing pattern intensity (PI);

performing gradient conversion (GC);

performing gradient difference (GD);

performing sum of local normalized correlation (SLNC);

performing variance-weighted sum of local normalized correlation (VWC); and

performing mutual information (MC).

16. The method of claim 9, wherein the step of transferring the result of the step of performing from the GPU to CPU, comprises:

transferring the result of the step of performing from an offline P-buffer to the CPU.

17. The method of claim 16, wherein the step of transferring the result of the step of performing from an offline P-buffer to the CPU, comprises:

rendering slices of the result of the step of performing;

transferring the slices from the offline P-buffer to the CPU; and

arranging the slices together to form a resulting image.

18. The method of claim 9, wherein the step of performing a registration application on the plurality of images using the GPU, comprises:

performing a registration application on the plurality of images using the GPU in parallel with using the CPU.

19. The method of claim 9, wherein the step of transferring a plurality of images from a central processing unit (CPU) to a graphics processing unit (GPU) comprises:

transferring a plurality of images from the CPU to a texture buffer of the GPU.

20. A program storage device readable by a machine, tangibly embodying a program of instructions executable on the machine to perform method steps for registering images, the method comprising the steps of:

transferring a plurality of images from a central processing unit (CPU) to a graphics processing unit (GPU);

performing a registration application on the plurality of images using the GPU; and

transferring the result of the step of performing from the GPU to CPU.