LMCS Reshaping Reduces Signaling Overhead in High Bit Depth Video Coding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

High bit depth in luma mapping chroma scaling (LMCS) based image/video coding leads to increased signaling overhead, necessitating a technique to reduce this overhead for efficient compression, transmission, and storage of high-resolution image/video data.

Innovation Solution

The method involves reshaping luma prediction sample values using reshaping related information that includes the absolute value and sign of a delta codeword to derive a linear or piecewise-linear reshaper, optimizing the signaling process and improving compression efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If high bit depth is used in LMCS based image coding, then image quality and resolution are improved, but signaling overhead increases

Engineering Contradiction:

Improveimage qualityVSAvoidsignaling overhead

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The piecewise-linear reshaper divides the luma sample value range into multiple segments or bins, applying different linear transformation parameters to each segment. This segmentation allows the complex high-bit-depth mapping to be represented through multiple simpler segments, reducing the signaling overhead while maintaining image quality.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The invention changes the parameter representation by using delta codewords with sign flags instead of directly signaling absolute transformation parameters. By encoding the difference from reference values and using sign bits to indicate direction, the number of bits required for signaling is reduced while preserving the ability to represent high-bit-depth transformations accurately.

Inventive Principle:

Principle #35Parameter changes

2Manufacturing precision

If piecewise-linear reshaper is used, then reshaping accuracy is improved, but signaling complexity increases

Engineering Contradiction:

Improvereshaping accuracyVSAvoidsignaling complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

Different linear transformation parameters are applied to different segments of the luma sample value range, allowing each segment to be optimized for its specific characteristics. This local optimization improves reshaping accuracy while the use of delta encoding and sign flags keeps the signaling complexity manageable by exploiting the local similarity between adjacent segments.

Inventive Principle:

Principle #3Local quality

3Loss of information

If delta codeword encoding is used, then signaling overhead is reduced, but decoding complexity increases

Engineering Contradiction:

Improvesignaling overheadVSAvoiddecoding complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

Reference values for the delta codewords are predetermined and stored in lookup tables during decoding. This preliminary preparation allows the decoder to efficiently retrieve reference values and perform simple addition operations with the received delta codewords and sign flags, reducing the actual decoding complexity despite the encoding scheme's sophistication.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11924431B2LMCS-based image coding method and device therefor

Publication Date: 2024.03.05 LG ELECTRONICS INC

AI summary

According to embodiment(s) disclosed in the present document, a reshaper model used in coding including LMCS can modify a value of a delta codeword on the basis of luma bit depth and perform reshaping on the basis of determination on whether or not a reshaper is linear, and thus, signaling overhead at a higher bit depth and overhead for signaling values of delta codewords for all bins can be reduced.