HDR Enhancement Layer Masking for Video Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current image and video coding technologies face challenges in efficiently delivering high dynamic range (HDR) content, particularly due to artifacts such as banding and blocking, which are exacerbated by the human visual system's non-linear response to light levels and the use of perceptual quantization, leading to suboptimal coding efficiency.

Innovation Solution

The proposed solution involves determining an optimal masking threshold for the enhancement layer in HDR coding, using a codeword range amplification method that restricts data coded by the enhancement layer and employs a non-linear quantizer, allowing for more efficient compression of image-like residual data, thereby reducing coding artifacts and improving visual quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If perceptual quantization is used in HDR coding to match human visual system response, then visual quality is improved, but coding efficiency deteriorates due to banding and blocking artifacts

Engineering Contradiction:
Improvevisual qualityVSAvoidcoding efficiency
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent divides the quantization process into two separate layers: base layer quantization and enhancement layer quantization. The base layer uses standard quantization while the enhancement layer applies additional quantization to correct artifacts. This segmentation allows each layer to be optimized for its specific function, resolving the contradiction between visual quality and coding efficiency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary enhancement layer that acts as a mediator between the base layer and the final decoded image. This enhancement layer contains residual information that compensates for quantization artifacts, effectively mediating the trade-off between compression efficiency and visual quality by adding corrective data without requiring complete re-quantization of the entire image.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Manufacturing precision

If dual-layer encoding is used for HDR content, then visual quality is improved, but bit rate efficiency deteriorates compared to single-layer encoding

Engineering Contradiction:
Improvevisual qualityVSAvoidbit rate
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent extracts only the essential residual information needed to correct quantization artifacts into the enhancement layer, rather than encoding complete duplicate images. By taking out only the necessary corrective data (quantization residuals and masking thresholds), the system achieves visual quality improvement without proportionally increasing bit rate, as the enhancement layer contains minimal essential information.

Inventive Principle:
Principle #2Taking out (Extraction)

3Loss of information

If enhancement layer codes all residual data, then compression completeness is improved, but coding complexity increases

Engineering Contradiction:
Improvecompression completenessVSAvoidcoding complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent applies local quality by using masking thresholds to determine which residual data should be coded in the enhancement layer. Instead of uniformly coding all residuals, the system selectively codes residuals only in regions where the masking threshold indicates they are visually significant. This local approach maintains compression completeness for important details while reducing coding complexity by ignoring insignificant regions.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3446484B1Enhancement layer masking for high-dynamic range video coding
Publication Date: 2023.09.06 DOLBY LABORATORIES LICENSING CORP
  • EP3446484B1 patent drawingFigure 1
  • EP3446484B1 patent drawingFigure 2A~2B
  • EP3446484B1 patent drawingFigure 3

AI summary

A method to improve the efficiency of coding high-dynamic range (HDR) signals in a dual-layer system is presented. A piece-wise linear, two-segment, inter-layer predictor is designed where base-layer codewords larger than a highlights threshold (Sh) are all mapped to a constant value. Given a target bit rate for the enhancement layer, which can be expressed as a percentage (β) of the bit rate of the base layer, an optimal highlights threshold is derived by computing estimated bit rates for the base and enhancement layers based on pixel complexity measures of pixels in the input HDR signal and the threshold value, and by minimizing an optimization criterion.