Compression of stereoscopic images

• US 6,205,241 B1
• Filed: 06/01/1998
• Issued: 03/20/2001
• Est. Priority Date: 06/01/1998
• Status: Expired due to Term

• Current Assignee: Canon Ayutthaya Limited (Canon Inc.)
• Original Assignee: Canon Ayutthaya Limited (Canon Inc.)
• Inventors: Melen, Roger D.
• Primary Examiner(s): Mehta, Bhavesh
• Assistant Examiner(s): Bayat, Ali
• Application Number:
  US09/088,617
• Time in Patent Office: 1,023 Days
• Field of Search: 382/100, 382/103, 382/106, 382/154, 382/239, 382/232, 382/276, 382/191, 382/284, 250/558, 356/12, 358/88, 358/107, 358/450, 364/561, 364/514, 348/47, 348/48, 348/42
• US Class Current: 382/154
• CPC Class Codes: G06T 9/005 Statistical coding, e.g. Huffman, run length coding

Abstract

A set of stereoscopic images (300) is compressed. For apparent points (404) in the set of stereoscopic images (300), a region (502) is identified in each image (300) which represents at least the apparent point (404). The locations of these regions (502) within the images (300), together with the geometry of the vantage point (306) locations, specify the apparent depths (416) of the apparent points (404) in the scene. Information relating to the apparent depths (416) is recorded for the apparent points (404). This recorded depth information, together with just one of the stereoscopic images (300), can be used to later reconstruct the other stereoscopic image (300) for stereoscopic viewing. The set of stereoscopic images (300) can be still images or moving images, and they can be captured digitally, scanned from photographs, or computer-generated.

Claims

1. A method for compressing one of a stereoscopic set of images, each of the images in the stereoscopic set of images having a vantage point associated therewith, the method comprising the steps of:

a) identifying in a first image of the stereoscopic set of images a first region representing an apparent point;

b) identifying in a second image of the stereoscopic set of images a second region representing the apparent point;

c) determining an epipolar offset, where the epipolar offset is the difference in an epipolar direction between the location of the second region relative to a characteristic point of the second image and the location of the first region relative to a characteristic point of the first image, which epipolar direction is substantially parallel to a vector from the vantage point of the first image to the vantage point of the second image; and

d) creating a compressed stereoscopic image by replacing the second region with an indication of the first region and with the corresponding epipolar offset.

2. The method of claim 1, wherein the step of identifying in a first image a first region representing an apparent point comprises the sub-steps of:

using contrast information in the first image to determine edge portions of the first image, which edge portions represent edges of an object; and

determining as the first region a portion of the first image which is generally circumscribed by the edge portions.

3. The method of claim 2, wherein the step of identifying in a second image a second region representing the apparent point comprises the sub-steps of:

using contrast information in the second image to determine edge portions of the second image, which edge portions represent edges of the object; and

determining as the second region a portion of the second image which is generally circumscribed by the edge portions.

4. The method of claim 1, wherein the step of identifying in a first image a first region representing an apparent point comprises selecting as the first region a portion of the first image, and the step of identifying in a second image a second region representing the apparent point comprises the sub-steps of:

selecting in the second image more than one target region, where each target region is the same size as the first region;

performing a cross-correlation calculation to determine the degree of similarity between the first region and each target region; and

selecting as the second region that target region which is most similar to the first region.

5. The method of claim 4, wherein the size of the first region is a predetermined number of pixels.

6. The method of claim 5, wherein the size of the first region is one pixel.

7. The method of claim 1, wherein steps (a) through (d) are repeated for a plurality of first and second regions and wherein the method further comprises the step of creating a compressed stereoscopic image by replacing the second image with a list of locations of the plurality of the first regions and the corresponding epipolar offsets.

8. The method of claim 1, wherein one of the first image and the second image is a scanned photograph of a physical scene.

9. The method of claim 1, wherein one of the first image and the second image is an image of a physical scene which has been captured in digital form.

10. The method of claim 1, wherein one of the first image and the second image is a computer generated image.

11. A method for using a first image from a stereoscopic set of images and a compressed form of a second image from the stereoscopic set to construct the second image in uncompressed form, which compressed form of the second image comprises information specifying an epipolar offset which is the difference in an epipolar direction between the location of a first region of the first image relative to a characteristic point of the first image and the location of a second region of the second image relative to a characteristic point of the second image, which epipolar direction is substantially perpendicular to the direction of view of the first image, the method comprising:

identifying the first region in the first image; and

producing a third image as an uncompressed form of the second image by overlaying the first region, displaced by the epipolar offset, on a duplicate of the first image.

12. The method of claim 11, further comprising the step of causing a background region of the duplicate of the first image to be replaced with a background pattern, where the background region is a portion of the duplicate of the first image which does correspond to the location of the undisplaced first region, and does not correspond to the location of the first region displaced by the epipolar offset.

13. The method of claim 12, wherein the background pattern comprises a combination of image qualities drawn from portions of the first image surrounding the first region.

14. The method of claim 12, wherein the background pattern is a solid color.

15. The method of claim 12, wherein the compressed form of the second image further comprises background image information based on the portion of the second image which corresponds to the location of the background region, and the background pattern is based on the background image information.

16. A computer apparatus for compressing one of a stereoscopic set of images, each of the images in the stereoscopic set of images having a vantage point associated therewith, the apparatus comprising:

a central processing unit (CPU);

an image memory, coupled to the CPU, for storing a first image of the stereoscopic set of images and a second image of the stereoscopic set of images; and

a program memory coupled to the CPU, for storing an array of instructions, which instructions, when executed by the CPU, cause the CPU to;

(a) identify in the first image a first region representing an apparent point;

(b) identify in the second image a second region representing the apparent point;

(c) determine an epipolar offset, where the epipolar offset is the difference in an epipolar direction between the location of the second region relative to a characteristic point of the second image and the location of the first region relative to a characteristic point of the first image, which epipolar direction is parallel to a vector from the vantage point of the first image to the vantage point of the second image; and

(d) create a compressed stereoscopic image by replacing the second region with an indication of the first region and with the corresponding epipolar offset.

17. The apparatus of claim 16, wherein the steps (a) through (d) are repeated for a plurality of first and second regions and wherein the method further comprises the step of creating a compressed stereoscopic image by replacing the second image with a list of locations of the plurality of the first regions and the corresponding epipolar offsets.

18. A computer readable medium containing a computer program for compressing one of a stereoscopic set of images, each of the images in the stereoscopic set of images having a vantage point associated therewith, the computer program performing the steps of:

(a) identifying in a first image of the stereoscopic set of images a first region representing an apparent point;

(b) identifying in a second image of the stereoscopic set of images a second region representing the apparent point;

(c) determining an epipolar offset, where the epipolar offset is the difference in an epipolar direction between the location of the second region relative to a characteristic point of the second image and the location of the first region relative to a characteristic point of the first image, which epipolar direction is substantially parallel to a vector from the vantage point of the first image to the vantage point of the second image; and

(d) creating a compressed stereoscopic image by replacing the second region with an indication of the first region and with the corresponding epipolar offset.

19. The computer readable medium of claim 18, wherein the steps (a) through (d) are repeated for a plurality of first and second regions and wherein the method further comprises the step of creating a compressed stereoscopic image by replacing the second image with a list of locations of the plurality of the first regions and the corresponding epipolar offsets.

20. A computer readable medium containing a computer program for using a first image from a stereoscopic set of images and a compressed form of a second image from the stereoscopic set to construct the second image in uncompressed form, which compressed form of the second image comprises information specifying an epipolar offset which is the difference in an epipolar direction between the location of a first region of the first image relative to a characteristic point of the first image and the location of a second region of the second image relative to a characteristic point of the second image, which epipolar direction is substantially perpendicular to the direction of view of the first image, the computer program performing the steps of:

identifying the first region in the first image; and

producing a third image as an uncompressed form of the second image by overlaying the first region, displaced by the epipolar offset, on a duplicate of the first image.

21. A method for compressing one of a stereoscopic set of images, the method comprising the steps of:

a) identifying in a first image of the stereoscopic set of images a first region representing an apparent point;

b) identifying in a second image of the stereoscopic set of images a second region representing the apparent point;

c) determining a predetermined offset, where the predetermined offset is the difference between the location of the second region relative to a characteristic point of the second image and the location of the first region relative to a characteristic point of the first image; and

d) compressing the stereoscopic image by generating the second region according to an indication of the first region and the predetermined offset.

22. The method of claim 21, wherein the step of identifying in a first image a first region representing an apparent point comprises the sub-steps of:

using contrast information in the first image to determine edge portions of the first image, which edge portions represent edges of an object; and

determining as the first region a portion of the first image which is generally circumscribed by the edge portions.

23. The method of claim 22, wherein the step of identifying in a second image a second region representing the apparent point comprises the sub-steps of:

using contrast information in the second image to determine edge portions of the second image, which edge portions represent edges of the object; and

determining as the second region a portion of the second image which is generally circumscribed by the edge portions.

24. The method of claim 21, wherein the step of identifying in a first image a first region representing an apparent point comprises selecting as the first region a portion of the first image, and the step of identifying in a second image a second region representing the apparent point comprises the sub-steps of:

selecting in the second image more than one target region, where each target region is the same size as the first region;

performing a cross-correlation calculation to determine the degree of similarity between the first region and each target region; and

selecting as the second region that target region which is most similar to the first region.

25. The method of claim 24, wherein the size of the first region is a predetermined number of pixels.

26. The method of claim 25, wherein the size of the first region is one pixel.

27. The method of claim 21, wherein steps (a) through (d) are repeated for a plurality of first and second regions and wherein the method further comprises the step of creating a compressed stereoscopic image by replacing the second image with a list of locations of the plurality of the first regions and the predetermined offsets.

28. The method of claim 21, wherein one of the first image and the second image is a scanned photograph of a physical scene.

29. The method of claim 21, wherein one of the first image and the second image is an image of a physical scene which has been captured in digital form.

30. The method of claim 21, wherein one of the first image and the second image is a computer generated image.