Lifting-based view compensated compression and remote visualization of volume rendered images

• US 8,126,279 B2
• Filed: 06/26/2008
• Issued: 02/28/2012
• Est. Priority Date: 11/19/2007
• Status: Active Grant

• Current Assignee: Siemens Healthineers AG (Siemens AG), Arizona Board of Regents (University of Arizona)
• Original Assignee: Siemens AG, University of Arizona
• Inventors: Marcellin, Michael W., Bilgin, Ali, Lalgudi, Hariharan G., Nadar, Mariappan S.
• Primary Examiner(s): Sherali, Ishrat I
• Application Number:
  US12/146,582
• Publication Number:
  US 20090129690A1
• Time in Patent Office: 1,342 Days
• Field of Search: 382/154, 382382-243, 37524001-24029, 345/419
• US Class Current: 382/232
• CPC Class Codes: G06T 15/08 Volume rendering
• CPC Class Codes: H04N 19/647 using significance based coding, e.g. Embedded Zerotrees of Wavelets [EZW] or Set Partitioning in Hierarchical Trees [SPIHT]

Abstract

A method for compressing 2D images includes determining a depth map for each of a plurality of sequential 2D images of a 3D volumetric image, determining coordinate transformations the 2D images based on the depth maps and a geometric relationship between the 3D volumetric image and each of the 2D image, performing a lifting-based view compensated wavelet transform on the 2D images using the coordinate transformations to generate a plurality of wavelet coefficients and compressing the wavelet coefficients and depth maps to generate a compressed representation of the 2D images.

Claims

1. A method for compressing 2D images, the method comprising:

determining a depth map for each of a plurality of sequential 2D images of a 3D volumetric image, each depth map including a plurality of depths that correspond to points in the corresponding 2D image, each depth corresponding to a depth of a point in the 3D volumetric image that is representative of an intensity of the point in the 2D frame;

determining coordinate transformations between the 2D images based on the depth maps and a geometric relationship between the 3D volumetric image and each of the 2D images;

performing a lifting-based view compensated wavelet transform on the 2D images using the coordinate transformations to generate a plurality of wavelet coefficients; and

compressing the wavelet coefficients and depth maps to generate a compressed representation of the 2D images.

2. The method of claim 1, wherein the geometric relationship comprises a distance, an elevation angle, and an azimuth angle between the 3D volumetric image and each of the 2D images.

3. The method of claim 1, wherein performing a lifting-based view compensated wavelet transform comprises:

generating high pass coefficient frames using a coordinate transformation from a first viewpoint to a prior viewpoint and a coordinate transformation from the first viewpoint to a future viewpoint; and

generating low pass coefficient frames using a coordinate transformation from a second viewpoint to a prior viewpoint and a coordinate transformation from the second viewpoint to a future viewpoint.

4. The method of claim 1, wherein each coordinate transformation comprises:

a first coordinate transformation from an origin viewpoint of the 3D volume to a viewpoint of a first one of the 2D images; and

a second coordinate transformation from a viewpoint of a second one of the 2D images to the origin viewpoint.

5. The method of claim 4, wherein the first and second coordinate transformations are respectively represented as a first matrix and a second matrix, and the coordinate transformation is determined by performing a matrix multiplication on the first and second matrixes.

6. The method of claim 1, wherein the 2D images are restored by:

decompressing the compressed representation of the 2D images to generate restored wavelet coefficients and restored depth maps; and

performing a lifting-based view compensated inverse wavelet transform on the restored wavelet coefficients using the restored depth maps.

7. The method of claim 1, further comprising rendering the 2D images from the 3D volumetric image using ray-casting.

8. A system for performing remote visualization of an image, the system comprising:

a network;

a server workstation comprising;

a depth map generation unit to determine a depth map for sequential 2D images of a 3D volumetric image;

a mapping unit to determine coordinate transformations between the 2D images based on the depth maps and a geometric relationship between the 3D volumetric image and each of the 2D images; and

a compressing unit to performing a lifting-based view compensated wavelet transform on the 2D images using the coordinate transformations to generate a plurality of wavelet coefficients and compress the wavelet coefficients and depth maps to generate a compressed representation of the 2D images;

a client workstation to receive the compressed representation of the 2D images across the network and to restore the 2D images from the compressed representation; and

a client display to display the restored 2D images.

9. The system of claim 8, wherein the client workstation comprises a decompressing unit to decompress the compressed representation of the 2D images to generate restored wavelet coefficients and restored depth maps.

10. The system of claim 9, wherein the client workstation decompressing unit is configured to perform a lifting-based view compensated inverse wavelet transform on the restored wavelet coefficients using the restored depth maps.

11. The system of claim 10, wherein the client workstation further comprises a graphical processing unit (GPU) to perform the inverse wavelet transform.

12. The system of claim 8, wherein the compressing unit comprises a JPEG2000 encoder and the compression of the wavelet coefficients and depth maps are performed using the JPEG2000 encoder.

13. The system of claim 9, wherein the decompressing unit comprises a JPEG2000 decoder and the decompression of the wavelet coefficients and depth maps are performing using the JPEG2000 decoder.

14. The system of claim 8, further comprising a hardware rendering engine to render the sequential 2D images from the 3D volumetric image.

15. The system of claim 14, wherein the hardware rendering engine renders the 2D images using ray casting.

16. The system of claim 15, wherein the depth for a point in the 2D frame is calculated using a sum of weighted depths of sample points along a corresponding ray cast from the ray casting.

17. The system of claim 16, wherein the depths of the sample points are weighted based on an amount each of the sample points contribute to the intensity of the point in the 2D frame.

18. The system of claim 8, further comprising a client input device to select desired viewpoints, wherein the client workstation sends the selected viewpoints across the network to the server workstation and the server workstation renders the 2D images based on the received viewpoints.

19. A non-transitory computer readable medium, tangibly embodying a program of instructions executable by the computer to perform method steps for compressing 2D images of a 3D volumetric image, the method comprising:

rendering the 2D images from the 3D volumetric data using ray casting;

determining a depth for each point of each 2D image based on a weighted sum of sample point depths in the 3D volumetric image, wherein each sample point is along a corresponding a ray cast by the ray casting;

determining coordinate mappings between the 2D images based on the depths and a geometric relationship between the 3D volumetric image and the 2D images;

performing a lifting-based view compensated wavelet transform on the 2D images using the coordinate mappings to generate a plurality of wavelet coefficients; and

compressing the wavelet coefficients and depths to generate a compressed representation of the 2D images.

20. The non-transitory computer readable medium of claim 19, wherein the geometric relationship comprises a distance, an elevation angle, and an azimuth angle between the 3D volumetric image and each of the 2D images.