Signed Distance Field Editing for Smooth 3D Anatomy Boundaries

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing volume editing and visualization systems in medical imaging suffer from low rendering quality due to interpolation artifacts at segmentation boundaries, particularly in complex anatomical structures with sparse blood vessels and low signal-to-noise ratios, limiting their effectiveness in clinical applications.

Innovation Solution

The use of signed distance fields (SDFs) for interactive editing and visualization, enabling real-time construction and fusion of SDFs from different sources, with direct and indirect rendering techniques to support editing workflows, allowing sub-voxel precision and smooth boundary reconstruction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If binary voxel map or label map is modified directly in existing volume editing systems, then editing operations can be performed, but interpolation artifacts appear at segmentation boundaries reducing rendering quality

Engineering Contradiction:
Improveediting operation capabilityVSAvoidsegmentation boundary quality
Core Design Contradiction:
Ease of operationVSManufacturing precision

Solution Approach 1:

The patent introduces signed distance fields (SDFs) as an intermediary data structure between binary voxel maps and final segmentation results. Instead of modifying binary voxel maps directly, the system performs editing operations on SDFs, which continuously represent spatial distance information. This intermediary representation allows smooth boundary modifications without creating interpolation artifacts, as the SDF naturally encodes boundary information in a differentiable manner that preserves rendering quality during interactive editing operations.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Manufacturing precision

If smoothed mask or antialiased SDF is derived from binary segmentation to reduce rendering artifacts, then rendering quality improves, but the algorithm becomes unsuitable for interactive editing

Engineering Contradiction:
Improverendering qualityVSAvoidinteractive editing speed
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent makes the signed distance field data structure serve dual purposes: it is both the working medium for interactive editing operations and the source for high-quality rendering. The SDF automatically provides smoothed boundaries and anti-aliasing effects during rendering without requiring separate post-processing steps. This self-service capability allows the same data structure to support both interactive editing speed and high rendering quality simultaneously, eliminating the trade-off between editing productivity and rendering precision.

Inventive Principle:
Principle #25Self-service

3Measurement precision

If sub-voxel precision editing is implemented using SDFs, then boundary accuracy improves, but computational complexity increases

Engineering Contradiction:
Improveboundary precisionVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces traditional mechanical voxel-based editing mechanisms with a field-based mathematical approach using signed distance fields. Instead of manipulating discrete binary voxels through complex interpolation algorithms, the system uses continuous distance field representations that naturally support sub-voxel precision through mathematical operations. This substitution of mechanical voxel manipulation with field-based computation simplifies the underlying complexity while achieving higher precision boundary representation and editing capabilities.

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

Data Source

PatentEP4033451B1Interactive image editing using signed distance fields
Publication Date: 2026.02.25 SIEMENS HEALTHINEERS AG
  • EP4033451B1 patent drawingFigure 1
  • EP4033451B1 patent drawingFigure 2
  • EP4033451B1 patent drawingFigure 3

AI summary

The present invention relates to a renderer (100) and an interactive method for image editing of medical 3D anatomical data. The method comprises receiving (S1) a dataset with volumetric image data, which have been acquired from an image acquisition modality (M) the received volumetric image data and providing (S2) a signed distance field data structure of the received dataset. Further, performing (S3) editing operations on a user interface (140) for editing at least a part of the provided signed distance field data structure and calculating (S5) a visualization of the editing operations and displaying (S6) the calculated visualization on a display (130) .