Medical Image Visualization Using 3D Mask Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing medical imaging techniques struggle to provide accurate and simplified three-dimensional representations of anatomical structures, as conventional methods like polygonal surface models fail to visualize the interior of organs and require significant data and can lead to inaccurate visualizations.

Innovation Solution

A method involving segmentation of medical image data into predetermined classes, generating a 3D mask, and applying a beam scanning technique to shift segmented volume elements based on a translation vector, allowing for the visualization of internal structures without intermediate representations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If polygonal surface models are used to represent anatomical details, then individual structures can be modeled and moved independently, but the interior of anatomical structures cannot be examined and significant data is required

Engineering Contradiction:
ImproveIndependent modeling and movement of structuresVSAvoidData requirements
Core Design Contradiction:
Ease of operationVSQuantity of substance

Solution Approach 1:

The volume data is segmented into different anatomical structures using classification algorithms. Each structure is assigned a unique class label, allowing independent manipulation while maintaining the original volumetric data integrity. This enables selective movement and examination of individual structures without requiring separate surface models for each.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from traditional 2D slice data to 3D volumetric representation with an additional classification dimension. By assigning class labels to voxels in three-dimensional space, the system enables both surface and interior examination simultaneously, eliminating the need for separate surface models.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Ease of operation

If polygonal surface models are used to represent anatomical details, then individual structures can be moved independently, but accurate visualization is compromised and intermediate representations are required

Engineering Contradiction:
ImproveIndependent movement of structuresVSAvoidVisualization accuracy
Core Design Contradiction:
Ease of operationVSManufacturing precision

Solution Approach 1:

The volume data is segmented into different anatomical structures using classification algorithms. Each structure is assigned a unique class label, allowing independent manipulation while maintaining the original volumetric data integrity. This enables selective movement and examination of individual structures without requiring separate surface models.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of creating intermediate surface model copies that lose interior information, the system works directly with the original volumetric data. Classification labels are applied to the volume voxels themselves, allowing accurate representation without loss of detail or creation of intermediate representations.

Inventive Principle:
Principle #26Copying

3Loss of information

If the entire volume data is visualized, then complete anatomical context is provided, but detailed examination of specific areas is difficult

Engineering Contradiction:
ImproveAnatomical contextVSAvoidDetailed examination of specific areas
Core Design Contradiction:
Loss of informationVSDifficulty of detecting and measuring

Solution Approach 1:

The system applies local quality by allowing different visualization parameters and levels of detail for different regions of the volume data. Specific areas can be highlighted, enhanced, or examined in detail while maintaining the overall anatomical context in the background, enabling both global overview and local detailed examination simultaneously.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The volume data is segmented into different anatomical structures using classification algorithms. Each structure is assigned a unique class label, allowing independent manipulation while maintaining the original volumetric data integrity. This enables selective movement and examination of individual structures without requiring separate surface models.

Inventive Principle:
Principle #1Segmentation

4Shape

If volume rendering is used to create three-dimensional representations, then anatomical structures are illustrated, but the data cannot be easily broken down into individual structures

Engineering Contradiction:
ImproveThree-dimensional representationVSAvoidData breakdown capability
Core Design Contradiction:
ShapeVSDevice complexity

Solution Approach 1:

The volume data is segmented into different anatomical structures using classification algorithms. Each structure is assigned a unique class label, allowing independent manipulation while maintaining the original volumetric data integrity. This enables selective movement and examination of individual structures without requiring separate surface models.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system provides dynamic control over the visualization, allowing users to selectively display, hide, or manipulate individual classified structures within the volume rendering. The classification framework enables flexible reconfiguration of the displayed anatomy without regenerating the entire volume model.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3712855B1Visualization of medical image data
Publication Date: 2026.01.07 SIEMENS HEALTHINEERS AG
  • EP3712855B1 patent drawingFigure 1
  • EP3712855B1 patent drawingFigure 2
  • EP3712855B1 patent drawingFigure 3

AI summary

A method for visualizing medical image data (BD) as volume data (VD) is described. In this method, medical image data (BD) is acquired. A 3D mask is generated by segmenting the image data (BD) and dividing the segmented areas (SG) into predefined classes. The image data (BD) and the mask data (MD) are then stored in two separate 3D texture files. A translation vector (Vec_tr) is then calculated, which describes the displacement of a segmented volume element (SG) between an origin position (pos_HF) and a target position (pos_HC). Furthermore, a visual representation of the image data (BD) is generated by applying a beam scanning method to the stored image data (BD). Finally, a displacement of a segmented volume element (SG) in the visual representation is performed by the translation vector (Vec_tr). A visualization device (50) is also described.Furthermore, a medical imaging system (60) is described.