Encoding and decoding of 3D HDR images using a tapestry representation

Patent metadata

Field	Value
Publication number	US-10469871-B2
Application number	US-201514964862-A
Country	US
Kind code	B2
Filing date	Dec 10, 2015
Priority date	Dec 18, 2014
Publication date	Nov 5, 2019
Grant date	Nov 5, 2019

Abstract

Representation and coding of multi-view images using tapestry encoding are described for standard and enhanced dynamic ranges compatibility. A tapestry comprises information on a tapestry image, a left-shift displacement map and a right-shift displacement map. Perspective images of a scene can be generated from the tapestry and the displacement maps. Different methods for achieving compatibility are described.

First claim

What is claimed is:

1. A computer-based method comprising:
providing an original first view image of a scene at a first dynamic range;
providing an original second view image, different than the original first view image, of the scene;
creating an original autostereoscopic or stereoscopic tapestry image at the first dynamic range by inserting pixels from the original second view image into a copy of the original first view image in accordance with an occlusion map, such that the original autostereoscopic or stereoscopic tapestry image comprises disoccluded regions containing pixels occluded in the first view image but non occluded in the original second view image;
generating a first view image at a second dynamic range, wherein the first dynamic range is higher than the second dynamic range;
predicting, by a computer and based on the original first view image, a first view image at the first dynamic range from the first view image at the second dynamic range, thereby obtaining a predicted first view image;
providing, by a computer, a first displacement map, wherein the first displacement map comprises distance information between the original first view image at the first dynamic range and the original autostereoscopic or stereoscopic tapestry image at the first dynamic range;
applying an inverse displacement operation to the predicted first view image, thereby obtaining a predicted tapestry image, wherein the inverse displacement operation comprises shifting pixels from a position in the predicted first view image to a position in the predicted tapestry image based on the first displacement map;
interpolating unfilled pixels in the predicted tapestry image; and
calculating a difference between the original autostereoscopic or stereoscopic tapestry image and the predicted tapestry image, thereby obtaining a residual.

2. The computer-based method of claim 1 , wherein the generating the first view image at the second dynamic range comprises a content mapping operation.

3. The computer-based method of claim 1 , wherein the predicting, by a computer, the first view image at the first dynamic range comprises an inverse mapping operation.

4. The computer-based method of claim 1 , further comprising:
compressing the first view image at the second dynamic range, thereby obtaining a compressed first view image at the second dynamic range; and
decompressing the compressed first view image, thereby obtaining a decompressed first view image,
wherein the predicting, by a computer, the first view image at the first dynamic range is from the decompressed first view image.

5. The computer-based method of claim 4 , further comprising:
encoding the residual;
transmitting the compressed first view image as a backward-compatible layer; and
transmitting the encoded residual as an enhancement layer.

6. The computer-based method of claim 5 , wherein the encoding the residual comprises a non-linear quantizer operation, a downsampling operation and a compression operation.

7. The computer-based method of claim 6 , wherein the predicting, by a computer, a first view image at the first dynamic range from the first view image at the second dynamic range comprises determining prediction parameters, and further comprising transmitting the prediction parameters as metadata.

8. The computer-based method of claim 6 , further comprising transmitting parameters from the non-linear quantizer operation as metadata.

9. The computer-based method of claim 1 , wherein the original autostereoscopic or stereoscopic tapestry image and the predicted tapestry image are frames of a video.

10. The computer-based method of claim 1 , wherein the applying an inverse displacement operation further comprises:
comparing a displacement for at least two pixels, thereby determining a pixel of the at least two pixels having a greater displacement than remaining pixels of the at least two pixels; and
selecting the pixel having the greater displacement.

11. The computer-based method of claim 1 , further comprising:
compressing the first displacement map, thereby obtaining a compressed first displacement map,
wherein the applying an inverse displacement operation to the predicted first view image is based on the compressed first displacement map.

12. The computer-based method of claim 11 , wherein the compressing the first displacement map comprises fitting a piecewise function with a piecewise polynomial.

13. A computer-based method comprising:
receiving a first view image at a first dynamic range;
receiving a second view image different than the first view image:
predicting, by a computer, a first view image at a second dynamic range from the first view image at the first dynamic range, thereby obtaining a predicted first view image, wherein the first dynamic range is lower than the second dynamic range;
receiving a first displacement map, wherein the first displacement map comprises distance information between an original first view image at the second dynamic range and an original autostereoscopic or stereoscopic tapestry image at the second dynamic range;
applying an inverse displacement operation to the predicted first view image, thereby obtaining a predicted tapestry image, wherein the inverse displacement operation comprises shifting pixels from a position in the predicted first view image to a position in the predicted tapestry image based on the first displacement map;
interpolating unfilled pixels in the predicted tapestry image;
receiving an enhancement layer that was built from a residual of a tapestry image and a predicted tapestry image, the tapestry image comprising inserted pixels from an original second view image into an original first view image in accordance with an occlusion map, such that the tapestry image comprises disoccluded regions containing pixels occluded in the first view image but non occluded in the original second view image;
adding the enhancement layer to the predicted tapestry image, thereby obtaining a reconstructed tapestry image:
receiving a second displacement map, wherein the second displacement map comprises distance information between an original second view image at the second dynamic range and the original autostereoscopic or stereoscopic tapestry image at the second dynamic range; and
generating a desired view image, based on the first and second displacement maps, and on the reconstructed tapestry image,
wherein the original autostereoscopic or stereoscopic tapestry image comprises disoccluded regions.

14. The computer-based method of claim 13 , further comprising, prior to the adding the enhancement layer to the predicted tapestry image, decompressing the enhancement layer, upsampling the enhancement layer and applying a non-linear dequantizer operation to the enhancement layer.

15. The computer-based method of claim 14 , wherein the applying a non-linear dequantizer operation comprises receiving non-linear quantizer parameters.

16. The computer-based method of claim 13 , wherein the predicting, by a computer, a first view image comprises receiving prediction parameters.

17. The computer-based method of claim 13 , wherein the first view is a left view, a leftmost view, a right view, a rightmost view, or a central view of a 3D image.

18. A non-transitory computer-readable storage medium having stored thereon computer-executable instruction for executing a method with a computer in accordance with claim 1 .

Assignees

• Dolby Laboratories Licensing Corp

Inventors

• Su Guan-Ming
• Ward Gregory John
• Kunkel Timo
• Daly Scott
• Hulyalkar Samir N

Classifications

• H04N13/161

  Encoding, multiplexing or demultiplexing different image signal components (for multi-view video sequence encoding H04N19/597) · CPC title

• H04N19/597 Primary

  specially adapted for multi-view video sequence encoding · CPC title

• H04N19/30

  using hierarchical techniques, e.g. scalability (H04N19/63 takes precedence) · CPC title

• H04N19/17

  the unit being an image region, e.g. an object · CPC title

• H04N2213/005

  Aspects relating to the "3D+depth" image format · CPC title