Layer decomposition in hierarchical VDR coding

Patent metadata

Field	Value
Publication number	US-9924171-B2
Application number	US-201615280822-A
Country	US
Kind code	B2
Filing date	Sep 29, 2016
Priority date	Nov 4, 2011
Publication date	Mar 20, 2018
Grant date	Mar 20, 2018

Abstract

Techniques use multiple lower bit depth (e.g., 8 bits) codecs to provide higher bit depth (e.g., 12+ bits) high dynamic range images from an upstream device to a downstream device. Multiple layers comprising a base layer and one or more enhancement layers may be used to carry video signals comprising image data compressed by lower bit depth encoders to a downstream device, wherein the base layer cannot be decoded and viewed on its own. Lower bit depth input image data to base layer processing may be generated from higher bit depth high dynamic range input image data via advanced quantization to minimize the volume of image data to be carried by enhancement layer video signals. The image data in the enhancement layer video signals may comprise residual values, quantization parameters, and mapping parameters based in part on a prediction method corresponding to a specific method used in the advanced quantization. Adaptive dynamic range adaptation techniques take into consideration special transition effects, such as fade-in and fade-outs, for improved coding performance.

First claim

The invention claimed is:

1. A method, comprising:
receiving an input visual dynamic range (VDR) image in a sequence of input images, wherein the input VDR image is of a first bit depth;
selecting a specific advanced quantization function from a plurality of available advanced quantization functions for mapping VDR image data of the first bit depth to base layer image data of a second bit depth lower than the first bit depth;
applying the specific advanced quantization function to the input VDR image to generate an input base layer image, wherein the input base layer image is of the second bit depth;
wherein the input base layer image as generated by the selected specific advanced quantization function is non-backward-compatible;
compressing image data derived from the input base layer image into a base layer (BL) video signal; and
compressing at least a portion of image data derived from the input VDR image into one or more enhancement layer (EL) video signals,
wherein the advanced quantization method is selected from the plurality of available advanced quantization functions based on one or more factors including minimizing an amount of image data to be encoded into the one or more EL video signals relative to the input VDR image.

2. The method as recited in claim 1 , further comprising:
decoding a BL image from the BL video signal, the BL image corresponding to the input BL image;
selecting a prediction method from one or more prediction methods;
generating a prediction image based at least in part on the BL image using the prediction method;
generating residual values based on the prediction image and the input VDR image;
applying non-linear quantization to the residual values to generate output EL image data, the residual values comprising higher bit depth values, and the output EL image data comprising lower bit depth values; and
compressing the output EL image data into the one or more EL video signals.

3. The method as recited in claim 2 , wherein the prediction method is selected based on a correspondence relationship between the advanced quantization method and the prediction method.

4. The method as recited in claim 1 , wherein the advanced quantization method comprises one or more of global quantization, linear quantization, linear stretching, curve-based quantization, probability-density-function (Pdf) optimized quantization, LLoyd-Max quantization, partition-based quantization, perceptual quantization, or cross-color channel/vector quantization.

5. The method as recited in claim 1 , wherein the sequence of input images comprises a second different VDR input image; and the method further comprising:
selecting a second different specific advanced quantization method from the plurality of available advanced quantization methods;
applying the second specific advanced quantization method to the second input VDR image to generate a second input BL image;
compressing second image data derived from the second input base layer image into the BL video signal; and
compressing at least a portion of image data derived from the second input VDR image into the one or more EL video signals.

6. The method as recited in claim 5 , further comprising:
decoding a second different BL image from the BL video signal, the second BL image corresponding to the second input BL image;
selecting a second different prediction method from the one or more prediction methods;
generating a second prediction image based at least in part on the second BL image using the second prediction method;
generating second different residual values based on the second prediction image and the second input VDR image;
applying non-linear quantization to the second residual values to generate second output EL image data, the second residual values comprising higher bit depth values, and the second output EL image data comprising lower bit depth values; and
compressing the output EL image data into the one or more EL video signals.

7. The method as recited in claim 1 , wherein the image data in the input BL image is compressed by a first 8 bit encoder in a VDR encoder into the BL video signal, and wherein the at least a portion of image data in the input VDR image is compressed by a second 8 bit encoder in the VDR encoder into the one or more EL video signals.

8. The method as recited in claim 1 , wherein a first image container is used to hold the image data derived from the input BL image, and wherein a second different image container is used to hold the at least a portion of image data in the input VDR image.

9. The method as recited in claim 8 , wherein at least one of the first image container and the second image container comprises pixel values in one or more channels in a color space.

10. The method as recited in claim 8 , wherein at least one of the first image container and the second image container is selected from a plurality of image containers associated with a plurality of sampling schemes, and wherein the plurality of sampling schemes comprises any of: a 4:4:4 sampling scheme, a 4:2:2 sampling scheme, or a 4:2:0 sampling scheme.

11. The method as recited in claim 1 , further comprising:
determining a specific profile for processing the input VDR image into the BL and EL video signals; and
performing one or more operations related to the specific profile in processing the input VDR image into the BL and EL video signals.

12. A method, comprising:
generating at least a portion of image data of a VDR image of a first bit depth, in a sequence of input images, based on one or more enhancement layer (EL) video signals;
generating a base layer (BL) image, of a second bit depth lower than the first bit depth, based on a BL video signal, the BL image was generated by applying a specific advanced quantization method to the VDR image, the specific advanced quantization method is selected from a plurality of available advanced quantization methods for mapping VDR image data of the first bit depth to base layer image data of the second bit depth;
wherein the BL image as generated by the selected specific advanced quantization function is non-backward-compatible; and
reconstructing a version of the VDR image based on the BL image and the at least a portion of image data of the VDR image, the version of the VDR image comprises reconstructed VDR image data of the first bit depth,
wherein the advanced quantization method was selected from the plurality of available advanced quantization methods based on one or more factors including minimizing an amount of image data to be derived from the one or more EL video signals relative to a source VDR image.

13. The method as recited in claim 12 , further comprising:
receiving prediction metadata including a set of mapping parameters;
determining a prediction method based on the prediction metadata;
generating a prediction image based at least in part on the BL image using the prediction method;
reconstructing the higher bit depth version of the VDR image by combining the prediction image with the at least a portion of image data derived from the one or more EL video signals.

14. The method as recited in claim 13 , wherein the prediction method corresponds to the advanced quantization method.

15. The method as recited in claim 13 , wherein the advanced quantization method comprises one or more of: global quantization, linear quantization, linear stretching, curve-based quantization, probability-density-function (Pdf) optimized quantization, LLoyd-Max quantization, partition-based quantization, perceptual quantization, or vector quantization.

16. Th

Assignees

• Dolby Laboratories Licensing Corp

Inventors

• Su Guan-Ming
• Qu Sheng
• Hulyalkar Samir N
• Chen Tao
• Gish Walter C
• Koepfer Hubert

Classifications

• H04N19/30 Primary

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

• H04N19/98

  Adaptive-dynamic-range coding [ADRC] · CPC title

• H04N19/103

  Selection of coding mode or of prediction mode · CPC title

• H04N19/124 Primary

  Quantisation · CPC title

• H04N19/126

  Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers · CPC title