Point Cloud Encoding with Additional Patches for Missed Points

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video codecs are inadequate for encoding and decoding 3D video content like point clouds, leading to inefficiencies in compression and transmission, particularly due to the introduction of artifacts such as cracks and holes in reconstructed point clouds during the 3D to 2D projection process.

Innovation Solution

The system employs an additional points patch to include missed points in the compression process, which are identified and stored separately to improve the reconstruction of 3D point clouds, reducing the appearance of cracks and holes by projecting the point cloud at different surface depths and using existing video codecs for compression.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If point clouds are compressed using existing video codecs during 3D to 2D projection, then compression efficiency is improved, but visual quality deteriorates due to artifacts such as cracks and holes

Engineering Contradiction:
Improvecompression efficiencyVSAvoidvisual quality
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent segments the point cloud data into multiple patches at different surface depths (near layer and far layer). By dividing the projection task into multiple depth-based segments, the system can process and compress each layer separately, reducing the artifacts caused by traditional single-layer projection while maintaining compression efficiency through structured organization of the segmented data

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a depth dimension to the traditional 2D projection by creating multiple patches at different surface depths (near and far layers). This transforms the conventional single 2D projection into a multi-layered 3D-to-2D projection approach, adding the depth dimension as a new parameter that enables better preservation of visual quality while maintaining compression efficiency

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Manufacturing precision

If point clouds are transmitted uncompressed, then visual quality is maintained, but bandwidth consumption increases

Engineering Contradiction:
Improvevisual qualityVSAvoidbandwidth
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent segments point cloud data into multiple depth-based patches (near and far layers) that can be compressed and transmitted separately. This segmentation enables efficient compression of each layer independently, reducing overall bandwidth consumption while maintaining visual quality through the structured multi-layer approach

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes the projection parameters by introducing multiple surface depths (near and far layers) instead of a single projection plane. This parameter change enables more efficient compression ratios while preserving visual quality, as the multi-depth approach captures more geometric information that can be compressed more effectively than traditional single-plane projection

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3794559B1Method and apparatus for processing point cloud
Publication Date: 2024.12.04 SAMSUNG ELECTRONICS CO LTD
  • EP3794559B1 patent drawingFigure 1~2
  • EP3794559B1 patent drawingFigure 3
  • EP3794559B1 patent drawingFigure 4A

AI summary

A method for encoding a point cloud comprises generating a first 2D frame representing a first attribute of a 3D point cloud and a second 2D frame representing a second attribute of the 3D point cloud, the first and second 2D frames including respective clusters of projected points from the 3D point cloud, detecting missed points of the 3D point cloud not being included in the first 2D frame, generating a first points patch representing the first attribute and a second points patch representing the second attribute based on a subset of the missed points, including the first points patch in the first 2D frame and the second points patch in the second 2D frame, encoding the first 2D frame including the first points patch and the second 2D frame including the second points patch to generate a compressed bitstream, and transmitting the compressed bitstream.