Dynamic Point Cloud Processing with 2D Geometry Filtering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing technologies face challenges in efficiently compressing dynamic point clouds while maintaining an acceptable quality of experience, as they often require significant bitrate or storage space, which is crucial for distributing immersive worlds and other applications.

Innovation Solution

A two-layer-based point cloud encoding structure is employed, comprising a base layer for lossy representation and an enhancement layer for higher quality, along with image-based encoding using existing video codecs to convert point cloud data into video sequences, and metadata for interpretation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If dynamic point clouds are compressed using existing technologies, then quality of experience is maintained, but bitrate and storage space requirements become excessive

Engineering Contradiction:
Improvequality of experienceVSAvoidbitrate and storage space
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent divides the point cloud data into multiple layers (base layer and enhancement layers) with different quality levels. The base layer provides essential information at lower bitrate, while enhancement layers add incremental quality improvements. This segmentation allows receivers to select appropriate layers based on available bandwidth and storage, resolving the contradiction between quality and resource consumption.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transforms point cloud data into alternative representations by changing parameters such as coordinate systems (e.g., transforming to spherical coordinates), sampling rates, and precision levels. This parameter transformation enables more efficient compression by matching the data representation to the specific application requirements, reducing bitrate while maintaining acceptable quality.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If point cloud data is distributed to end-users, then immersive world distribution is enabled, but consumption of bitrate and storage space increases

Engineering Contradiction:
Improvedistribution capabilityVSAvoidbitrate and storage space
Core Design Contradiction:
Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

The patent implements dynamic adaptation of point cloud data distribution by adjusting the level of detail, resolution, and quality based on receiver capabilities, network conditions, and storage availability. The system can dynamically select which layers to transmit and at what quality level, enabling versatile distribution across different platforms while optimizing bitrate consumption according to actual needs.

Inventive Principle:
Principle #15Dynamics

3Reliability

If static colored huge point clouds are used for culture heritage and topography, then preservation and visualization are achieved, but data size becomes unmanageable

Engineering Contradiction:
Improvepreservation qualityVSAvoiddata size
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent extracts and separates different attributes of point cloud data (geometry, color, texture, semantic information) into independent layers or representations. This extraction allows selective compression and transmission of only the essential preservation-critical attributes, reducing overall data size while maintaining reliability for heritage and topography applications.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3942828B1Processing a point cloud
Publication Date: 2025.10.29 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP3942828B1 patent drawingFigure 1~2
  • EP3942828B1 patent drawingFigure 3
  • EP3942828B1 patent drawingFigure 3a

AI summary

At least one embodiment relates to a method for smoothing (filtering) the geometry of a point cloud frame by performing an analysis and filtering of said geometry point cloud in a 2D space, without reconstruction of 3D samples in a 3D space, and by using a flexible filtering operator that, in addition to moving existing points, can also remove points or add new ones.