Visual File Identifiers Using Content-Independent Semantic Metadata

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems struggle to determine whether two visual representations convey the same overall meaning or signified, especially when changes in content details occur, lacking effective identifiers that reflect the underlying semantic consistency or divergence.

Innovation Solution

Generate identifiers for visual representations using content-independent semantic metadata elements, employing rulesets that prioritize properties not directly contributing to the visual representation's meaning, ensuring that visual representations with the same signified receive the same identifier, even with varying content details.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If identifiers are generated using content-dependent metadata elements, then the identifier reflects specific visual details, but visual representations with the same overall meaning receive different identifiers

Engineering Contradiction:
Improveidentifier accuracyVSAvoidsemantic consistency
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent segments metadata elements into two distinct categories: content-independent semantic metadata elements (e.g., file format, creation software, dimensional properties) and content-dependent metadata elements (e.g., specific visual content, text labels). By generating identifiers using only the content-independent segment, the system maintains semantic consistency across visual representations with the same overall meaning while ignoring variations in specific visual details.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts and isolates the content-independent semantic metadata elements from the complete set of metadata associated with visual representations. This extraction process removes the problematic content-dependent elements that cause identifiers to differ unnecessarily, retaining only the stable semantic properties that truly reflect the overall meaning of the visual representation.

Inventive Principle:
Principle #2Taking out (Extraction)

2Loss of information

If identifiers are generated using all metadata elements, then comprehensive information is captured, but changes in non-semantic details create false differences

Engineering Contradiction:
Improvemetadata completenessVSAvoididentifier reliability
Core Design Contradiction:
Loss of informationVSReliability

Solution Approach 1:

The patent extracts only the relevant content-independent semantic metadata elements from the complete metadata set, discarding content-dependent elements that do not contribute to semantic meaning. This selective extraction preserves the reliability of identifiers by ensuring they reflect only meaningful semantic properties while maintaining sufficient information for accurate identification.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies different quality requirements to different metadata elements: content-independent semantic metadata elements are treated as high-quality, reliable indicators of overall meaning, while content-dependent metadata elements are treated as variable, lower-quality indicators that may change without affecting semantic equivalence. This local quality differentiation ensures identifier reliability.

Inventive Principle:
Principle #3Local quality

3Adaptability or versatility

If content-independent semantic metadata elements are used for identifier generation, then semantic consistency is maintained, but some visual representation properties are not reflected

Engineering Contradiction:
Improvesemantic consistencyVSAvoididentifier specificity
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent segments the identifier generation process into two distinct phases: first, generating a semantic identifier using content-independent metadata elements to ensure semantic consistency; second, optionally maintaining separate content-dependent metadata for applications that require knowledge of specific visual details. This segmentation allows the system to prioritize semantic consistency while preserving access to detailed information when needed.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4697204A1Visual file identifiers using content-independent semantic metadata elements
Publication Date: 2026.02.18 SAP SE
  • EP4697204A1 patent drawingFigure 1
  • EP4697204A1 patent drawingFigure 2A
  • EP4697204A1 patent drawingFigure 2B

AI summary

Visual representations represented in computer files can undergo various changes, in terms of specific visual details in a visual representation, without changing the semantics of what is represented. The present disclosure provides techniques and solutions for generating an identifier for a visual representation using one or more content-independent semantic metadata elements. Content-independent semantic properties are those that do not correspond to individual visual elements in a visual representation, or to properties of a visual representation that do not contribute to a description of an overall semantic provided by the visual representation. Disclosed techniques can be used to identify when a visual representation is generated or imported that has the same overall semantic, as indicated by the identifier, as another visual representation. The two visual representations can have the same or different image elements. Disclosed techniques also provide that visual representations that have a different overall semantic are assigned different identifiers.