Selective Volume Rendering Using Anatomical Connectivity

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Higher spatial resolution in medical images, particularly in CT scans, leads to cluttered and occluded visualization of less relevant anatomical structures, making it difficult to clearly visualize structures of interest for diagnosis.

Innovation Solution

A computer-implemented method for selective volume rendering using directional connectivity, which identifies and renders only the voxels connected to a starting point anatomy, such as the aorta, based on alignment and connectivity values, without requiring manual selection or machine learning-based segmentation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If comprehensive volume rendering of the entire volume of interest is performed, then complete anatomical structures are visualized, but the visualization becomes cluttered and occluded with less relevant structures

Engineering Contradiction:
Improvecompleteness of anatomical visualizationVSAvoidclutter and occlusion in visualization
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent segments the volume of interest by computing connectivity values for each voxel relative to a starting point anatomy (e.g., aorta). This segmentation identifies and separates relevant anatomical structures (connected to the starting point) from less relevant structures, enabling selective rendering that maintains completeness while reducing clutter.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by differentiating the rendering treatment of different voxel regions. Voxels with high connectivity values (anatomically connected to the starting point) are rendered with full visibility, while voxels with low connectivity values (less relevant structures) are rendered with reduced visibility or excluded, creating a visualization with varying local quality characteristics.

Inventive Principle:
Principle #3Local quality

2Object-affected harmful factors

If manual selection or machine learning-based segmentation is used to identify structures of interest, then relevant anatomy can be selectively visualized, but the process requires extensive training data and manual intervention

Engineering Contradiction:
Improveclutter and occlusion in visualizationVSAvoidrequirement for manual intervention and training data
Core Design Contradiction:
Object-affected harmful factorsVSDevice complexity

Solution Approach 1:

The patent implements self-service by enabling the system to automatically identify and segment relevant anatomical structures through computational connectivity analysis. The method computes alignment values and connectivity values based on the anatomical relationships within the volume itself, without requiring external training data or manual annotation, allowing the system to serve its own segmentation needs.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical/manual process of structure selection and machine learning-based segmentation with an automated computational approach. Instead of manual intervention or complex ML models requiring training data, the system uses mathematical computations of alignment and connectivity values to automatically identify and segment relevant anatomy.

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

Data Source

PatentEP4700709A1Selective volume rendering
Publication Date: 2026.02.25 KONINKLIJKE PHILIPS NV
  • EP4700709A1 patent drawingFigure 1a~1c
  • EP4700709A1 patent drawingFigure 2a~2b
  • EP4700709A1 patent drawingFigure 3a~3b

AI summary

The present disclosure concerns a computer-implemented method for volume rendering of a region comprising a structure in a volume of interest. The method comprises receiving data representing a plurality of voxels of the volume of interest representative of the region. The method comprises obtaining a starting point voxel representative of a starting point anatomy which is connected to the structure. The method comprises determining an alignment value indicative of a direction alignment of the respective voxel relative to neighbored voxels in a neighborhood of the respective voxel. The method comprises determining, based on the alignment value per voxel, a connectivity value per voxel of the plurality of voxels relative to the starting point voxel, wherein the connectivity value is indicative of an anatomical connection relative to the starting point anatomy. The method comprises identifying, based on the connectivity value, voxels of interest representing connected structures in the region comprising the structure. The method comprises volume rendering of the voxels of interest for visualizing the structure.