Avatar Landmark Signaling in 3D Scene Description Formats

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current scene description formats lack a method to signal landmark sets for avatars, necessitating separate files for avatar models and landmark information, which complicates interoperability and feature implementation in extended reality applications.

Innovation Solution

The proposed solution involves extending scene description formats like glTF and AJIF to include explicit signaling of landmark sets, using vertex indices, face indices, and barycentric weights to define landmarks, along with semantic types and descriptions, ensuring compatibility and ease of feature integration.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If separate files are used for avatar models and landmark information, then data organization is simplified, but interoperability and feature implementation become complicated

Engineering Contradiction:
Improvedata organizationVSAvoidinteroperability
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent merges previously separate avatar model data and landmark information into a unified scene description format. The landmark set information is embedded within the avatar model structure, allowing applications to access both mesh and landmark data from a single file, thereby improving interoperability while maintaining organized data structure through hierarchical nesting.

Inventive Principle:
Principle #5Merging (Combining)

2Device complexity

If landmark sets are not explicitly signaled in scene description formats, then format simplicity is maintained, but avatar feature implementation becomes difficult

Engineering Contradiction:
Improveformat structureVSAvoidfeature implementation
Core Design Contradiction:
Device complexityVSAdaptability or versatility

Solution Approach 1:

The patent segments landmark information into distinct structured fields (landmark set ID, type, description, and individual landmark definitions with vertex/face indices and barycentric weights) within the scene description format. This segmentation allows applications to efficiently access specific landmark data needed for different features while maintaining an organized, scalable structure that doesn't overly complicate the overall format.

Inventive Principle:
Principle #1Segmentation

3Ease of manufacture

If multiple files are required for complete avatar information, then data separation is achieved, but application setup and processing become more complex

Engineering Contradiction:
Improvedata separationVSAvoidapplication setup
Core Design Contradiction:
Ease of manufactureVSEase of operation

Solution Approach 1:

The patent combines multiple data sources (avatar mesh, landmark positions, semantic types, and descriptions) into a single integrated scene description file. This eliminates the need for applications to load and correlate multiple separate files, significantly simplifying application setup and processing while maintaining clear data separation through structured hierarchical organization within the unified format.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP4686211A1Signaling avatar landmarks in scene description
Publication Date: 2026.01.28 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP4686211A1 patent drawingFigure 1A
  • EP4686211A1 patent drawingFigure 1B
  • EP4686211A1 patent drawingFigure 1C

AI summary

Some embodiments of a method may include: obtaining information corresponding to an avatar, wherein the information comprises a set of attributes associated with the avatar; parsing the information for a first attribute of the set of attributes, wherein the first attribute comprises a connection of a mesh to the avatar; parsing the information for a second attribute of the set of attributes, wherein the second attribute corresponds to a type property of the avatar; and providing a feature related to the mesh in a 3D scene using the parsed attributes of the avatar.