Digital Image Encoding Using Persistent Homology for Quality Compression

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing digital image encoding methods, such as triangulation-based compression, face challenges in optimizing triangle placement to minimize image degradation while achieving efficient compression, particularly when decoded by modern computers with graphics processing capabilities.

Innovation Solution

An encoding method utilizing algorithmic topology principles, specifically persistent homology, to identify and retain topologically significant structures in a digital image, forming triangulation data from vertices with long lifetimes, and using Delaunay triangulation for reconstruction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If triangulation encoding is used to compress digital images, then compression efficiency is improved, but image quality degradation increases

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

Solution Approach 1:

The patent applies local quality by differentiating the treatment of vertices based on their topological persistence. Vertices with long lifetimes (high persistence) are retained to maintain important structural information, while vertices with short lifetimes are discarded for compression. This selective retention strategy optimizes the balance between compression efficiency and image quality by focusing on topologically significant features rather than uniformly treating all vertices equally.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent changes the parameter of vertex selection from arbitrary or uniform sampling to persistence-based selection. By using persistent homology to compute lifetime values for each vertex and selecting vertices based on their persistence thresholds, the method transforms the vertex selection process into a parameter-driven optimization that balances compression ratio and reconstruction quality.

Inventive Principle:
Principle #35Parameter changes

2Manufacturing precision

If more triangles are used to represent the image, then image reconstruction quality is improved, but data size increases

Engineering Contradiction:
Improveimage reconstruction qualityVSAvoiddata size
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent extracts only the essential topological information from the complete vertex set by using persistent homology to identify and retain only the vertices with long lifetimes. This extraction process removes redundant vertices that contribute minimally to the topological structure, resulting in a compressed representation that maintains image reconstruction quality while significantly reducing the number of triangles and data size required.

Inventive Principle:
Principle #2Taking out (Extraction)

3Ease of manufacture

If systematic or stochastic methods are used for triangle encoding, then encoding process is simplified, but optimization of triangle placement is insufficient

Engineering Contradiction:
Improveencoding process simplicityVSAvoidtriangle placement optimization
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The patent replaces mechanical or heuristic triangle placement methods with a topological field-based approach. Instead of using systematic grids or stochastic sampling, the method uses persistent homology to compute topological persistence values for all vertices, creating a data-driven selection mechanism that automatically identifies optimal vertices for retention based on their topological significance rather than positional or random criteria.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP4078959B1Method for encoding a digital image in order to compress same
Publication Date: 2025.07.30 SHADOW
  • EP4078959B1 patent drawingFigure 1~2
  • EP4078959B1 patent drawingFigure 3a~3b
  • EP4078959B1 patent drawingFigure 3c~3d

AI summary

The invention relates to a method of encoding a digital image in order to compress same, the digital image being defined as a point cloud associating a set of N pixels, designated as vertices, to a scalar intensity value. The method aims at establishing triangulation points of the digital image and implements the principles of algorithmic topology.