Hybrid reference picture reconstruction method for single and multiple layered video coding systems

Patent metadata

Field	Value
Publication number	US-9756353-B2
Application number	US-201214370400-A
Country	US
Kind code	B2
Filing date	Dec 20, 2012
Priority date	Jan 9, 2012
Publication date	Sep 5, 2017
Grant date	Sep 5, 2017

Abstract

An inter-layer reference picture is generated either by considering an inverse mapped signal from a base layer, a temporal signal from an enhancement layer, or a combination of both.

First claim

The invention claimed is:

1. A method of generating a picture region, the method comprising:
providing a first layer comprising a plurality of first layer pictures having a first dynamic range associated with a first bit depth of the first layer pictures;
providing a second layer comprising a plurality of second layer pictures having a second dynamic range associated with a second bit depth of the second layer pictures; wherein the second bit depth is higher than the first bit depth;
performing a first mode of picture region generation comprising inverse mapping first bit depth data associated to the first dynamic range to second bit depth data associated to the second dynamic range of a region of a first layer picture from the plurality of first layer pictures to generate a first intermediate picture region;
performing a second mode of picture region generation comprising utilizing a second layer temporal signal from one or more previously decoded pictures in the second layer to generate a second intermediate picture region, the second layer temporal signal comprising one or more pixels from the one or more previously decoded pictures in the second layer, wherein the first intermediate picture region and the second intermediate picture region correspond to the picture region to be generated;
performing a third mode of picture region generation comprising, for each pixel location in the first intermediate picture region and the second intermediate picture region, computing an average of a pixel value corresponding to the pixel location in the first intermediate picture region and a pixel value corresponding to the same pixel location in the second intermediate picture region to generate a third intermediate picture region; and
selecting as the picture region to be generated one picture region from among the first intermediate picture region, the second intermediate picture region, and the third intermediate picture region based on a metric;
wherein selecting comprises:
determining one or more ranges based on characteristics of the inverse mapping from the first dynamic range to the second dynamic range of prior first and second layer pictures, wherein each range is bounded by a pair of threshold pixel values at the first bit depth;
determining whether or not a mean or other statistic of the region of the picture from the first layer is within one of the determined ranges;
and if the mean or other statistic of the region of the picture from the first layer is within one of the determined ranges: selecting the first intermediate picture region as the picture region to be generated, otherwise selecting the second intermediate picture region as the picture region to be generated.

2. The method according to claim 1 , wherein the performing of the first mode of picture region generation comprises:
utilizing a first layer temporal signal from one or more previously decoded pictures in the first layer; and
inverse mapping from the first dynamic range to the second dynamic range the first layer temporal signal to generate the first intermediate picture region,
wherein the first layer temporal signal comprises the region of the first layer picture.

3. The method according to claim 2 , wherein the utilizing of the first layer temporal signal from the one or more previously decoded pictures in the first layer comprises:
providing a motion vector associated with the first layer temporal signal;
applying the motion vector to the first layer temporal signal to generate a motion compensated first layer temporal signal at a certain position; and
copying the motion compensated first layer temporal signal at the certain position to a collocated position in the picture region.

4. The method according to claim 1 , wherein the utilizing of the second layer temporal signal from the one or more previously decoded pictures in the second layer comprises:
copying the second layer temporal signal from a collocated position in the one or more previously decoded pictures in the second layer to the collocated position in the picture region, wherein a motion vector associated with the second layer temporal signal comprises a zero vector.

5. The method according to claim 1 , wherein the utilizing of the second layer temporal signal from the one or more previously decoded pictures in the second layer comprises:
providing a motion vector associated with the second layer temporal signal;
applying the motion vector to the second layer temporal signal to generate a motion compensated second layer temporal signal at a certain position; and
copying the motion compensated second layer temporal signal at the certain position to a collocated position in the picture region.

6. The method according to claim 5 , further comprising explicitly signaling the motion vector at full precision to a decoding method.

7. The method according to claim 5 , further comprising explicitly signaling the motion vector at reduced precision to a decoding method.

8. The method according to claim 5 , wherein the motion vector associated with the second layer temporal signal is generated based on a collocated motion vector from the first layer.

9. The method according to claim 8 , wherein the providing of the motion vector associated with the second layer temporal signal comprises:
scaling the collocated motion vector from the first layer based on a relative resolution ratio between the first layer and the second layer.

10. The method according to claim 1 , wherein the selecting comprises selecting the third intermediate picture region as the picture region to be generated.

11. The method according to claim 1 , wherein the determining of whether or not a mean or other statistic of the region of the picture from the first layer is within one of the determined ranges comprises:
computing a mean of the pixel values in the region of the picture from the first layer; and
if the computed mean is within one of the pairs of threshold pixel values: determining that the region of the picture from the first layer is within one of the determined ranges.

12. The method according to claim 11 , wherein the pairs of threshold pixel values are determined prior to generating the region of the picture region.

13. The method according to claim 11 , wherein the pairs of threshold pixel values are determined by analyzing data from a sequence of pictures, a scene, or a group of pictures.

14. The method according to claim 1 , wherein the determining of whether or not a mean or other statistic of the region of the picture from the first layer is within one of the determined ranges comprises:
computing a mean of the pixel values in the region of the picture from the first layer; and
if the computed mean is within one of the pairs of threshold pixel values:
determining that the region of the picture from the first layer is within one of the determined ranges;

else:
determining that the region of the picture from the first layer is outside the determined ranges.

15. The method according to claim 14 , wherein the pairs of threshold pixel values are determined prior to generating the region of the picture region.

16. The method according to claim 14 , wherein the pairs of threshold pixel values are determined by analyzing data from a sequence of pictures, a scene, or a group of pictures.

17. A method of generating a picture region, the method comprising:
providing a first layer comprising a plurality of first layer pictures having a first dynamic range associated with a first bit depth of the first layer pictures;
providing a second

Assignees

• Dolby Laboratories Licensing Corp

Inventors

• Ye Yan
• Yin Peng
• He Yuwen
• Gish Walter C

Classifications

• H04N19/36 Primary

  Scalability techniques involving formatting the layers as a function of picture distortion after decoding, e.g. signal-to-noise [SNR] scalability · CPC title

• H04N19/103

  Selection of coding mode or of prediction mode · CPC title

• H04N19/10

  using adaptive coding · CPC title

• H04N19/187

  the unit being a scalable video layer · CPC title

• H04N19/55

  Motion estimation with spatial constraints, e.g. at image or region borders · CPC title