Point Cloud Quantization Step Parameter for Bitrate Distribution

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current point cloud compression techniques, such as TMC2, face limitations in efficiently managing bit allocation and rate distortion, particularly in dynamic 3D scenes, where auxiliary information encoding and prediction methods are restricted to immediate neighbors, leading to suboptimal bitrate distribution between luminance and chrominance.

Innovation Solution

The proposed solution involves parameterizing the quantization scheme using fixed-point arithmetic to calculate quantization step sizes based on user-input quantization parameters, allowing for different quantization parameters for chrominance and luminance, and enabling a finer rate-distortion optimization framework, with a table mapping quantization parameters to step sizes, thereby improving bit allocation between attribute channels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If the same quantization parameter is used for both chrominance and luminance, then the encoding process is simpler, but the bitrate distribution is suboptimal and compression efficiency is reduced

Engineering Contradiction:
Improveencoding process simplicityVSAvoidcompression efficiency
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The patent segments the quantization parameter control into separate parameters for luminance (QPL) and chrominance (QPC), allowing independent optimization of bitrate allocation for each color component. This segmentation enables fine-grained control over compression efficiency while maintaining manageable complexity through the established relationship QPC = QPL + Offset.

Inventive Principle:
Principle #1Segmentation

2Productivity

If different quantization parameters are used for chrominance and luminance, then bitrate distribution is optimized, but the device complexity increases

Engineering Contradiction:
Improvebitrate distribution optimizationVSAvoidquantization parameter management
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces a systematic parameter change approach where QPC is derived from QPL through a defined relationship (QPC = QPL + Offset). This allows the system to use different effective quantization parameters for optimized bitrate distribution while reducing device complexity by eliminating the need to independently manage completely separate parameter sets.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The Offset parameter acts as an intermediary that mediates between the luminance quantization parameter (QPL) and chrominance quantization parameter (QPC). This intermediary simplifies the management of different quantization parameters by providing a single control mechanism (the Offset) that adjusts the relationship between QPL and QPC, thereby optimizing bitrate distribution without significantly increasing device complexity.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Speed

If patch matching is limited to immediate neighbors only, then the prediction process is faster, but the coding efficiency of auxiliary information is reduced

Engineering Contradiction:
Improveprediction process speedVSAvoidcoding efficiency
Core Design Contradiction:
SpeedVSProductivity

Solution Approach 1:

The patent applies dynamics by making the prediction process adaptive rather than static. Instead of being limited to fixed immediate neighbors, the system dynamically selects reference patches from a broader search space based on matching criteria. This dynamic approach improves coding efficiency while maintaining reasonable speed by only expanding the search when beneficial.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3922032B1Quantization step parameter for point cloud compression
Publication Date: 2024.06.12 SONY GROUP CORP
  • EP3922032B1 patent drawingFigure 1
  • EP3922032B1 patent drawingFigure 2
  • EP3922032B1 patent drawingFigure 3

AI summary

A technique to parameterize the quantization scheme of the attribute coding of point cloud compression algorithms is described herein. Based on fixed-point arithmetic, the algorithm calculates the quantization step size (QS) in fixed-point notation, given a user-input quantization parameter (QP).