Split Point Compression Profiles for AI Point Cloud Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video coding systems, including wavelet-based, object-based, and block-based systems, are inadequate for effective compression and processing of point clouds.

Innovation Solution

Implementing split point configurations with compression profiles, where endpoints negotiate and select split point profiles that match artificial intelligence application requirements, determining compression formats, and processing intermediate data accordingly.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If current video coding systems (wavelet-based, object-based, block-based) are used for point cloud compression, then general video compression is achieved, but compression efficiency for point clouds is inadequate

Engineering Contradiction:
Improvecompression efficiencyVSAvoidadequacy for point cloud processing
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent introduces split point configuration profiles that define specific parameter settings for different AI application tasks. These profiles adjust compression parameters (such as precision, data format, and processing depth) to optimize the balance between compression efficiency and AI task performance requirements, resolving the inadequacy of general video coding systems for point cloud-specific applications.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system dynamically selects and negotiates split point profiles between endpoints based on real-time AI application requirements. This dynamic adaptation allows the compression system to adjust its characteristics according to specific task demands, improving both compression efficiency and adequacy for different point cloud processing scenarios.

Inventive Principle:
Principle #15Dynamics

2Productivity

If split point profiles are negotiated and selected for specific AI tasks, then compression is optimized for AI applications, but system complexity increases

Engineering Contradiction:
ImproveAI task processing efficiencyVSAvoidprofile negotiation and selection mechanism
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent pre-defines multiple split point configuration profiles with optimized parameters for different AI application types. These profiles are established in advance and stored in the system, eliminating the need for complex real-time optimization calculations. The negotiation process simply selects from these pre-configured options, reducing system complexity while maintaining AI task optimization.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

Different split point profiles are tailored to specific AI application requirements (e.g., classification, detection, segmentation tasks). Each profile contains locally optimized parameters suited for its target application, allowing the system to achieve high AI processing efficiency without requiring a single complex universal configuration mechanism.

Inventive Principle:
Principle #3Local quality

3Reliability

If compression formats are determined and intermediate data is processed according to split point profiles, then data transmission quality is improved, but processing time increases

Engineering Contradiction:
Improvedata transmission qualityVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The split point profiles define specific parameter settings for intermediate data processing, including compression format selection and data characteristic adjustments. By changing these parameters according to pre-defined profiles rather than performing complex real-time analysis, the system improves data transmission quality while minimizing additional processing time.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4651470A1Split points configurations with compression profile
Publication Date: 2025.11.19 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP4651470A1 patent drawingFigure 1A
  • EP4651470A1 patent drawingFigure 1B
  • EP4651470A1 patent drawingFigure 1C

AI summary

Systems, methods, and instrumentalities are disclosed associated with split point configurations with compression profile. For example, endpoints may select and/or negotiate split point configuration information (e.g., profiles). Split point configuration profiles may include a split point data characteristic and/or related compression characteristics that may be performed with an associated compression profile. Endpoints (e.g., a first endpoint and a second endpoint) may compute split point data characteristics and/or related compression characteristics of different split point profiles, for example, to select the split point profile that matches the expected application task (e.g., artificial intelligence (Al)) take requirements.