Vessel Graph Aggregation Across Frames With Missing Segment Alignment

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing medical imaging technologies face challenges in providing a consistent representation of vessel structures across multiple frames due to inconsistencies caused by patient motion, cardiac motion, and varying imaging angles, leading to reduced reliability and accuracy in downstream tasks.

Innovation Solution

A computer-implemented method that generates an aggregated representation of a vessel structure by attaching segments of one graph representation to another if the latter lacks corresponding segments, using spatial coordinates to align and augment graph representations from consecutive medical images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If graph representations are extracted from multiple frames to capture vessel structure, then the understanding of anatomy and physiology is improved, but inconsistencies arise due to patient motion, cardiac motion, and changing angulation

Engineering Contradiction:

Improvereliability of vessel structure representationVSAvoidconsistency of vessel structure representation

Core Design Contradiction:

ReliabilityVSStability of the object's composition

Solution Approach 1:

The patent merges multiple graph representations from different frames into a single aggregated graph representation. This is achieved by iterating through all graph representations, comparing segments between frames, and combining consistent segments while resolving inconsistencies through spatial coordinate matching and segment attachment operations, thereby producing a unified and reliable vessel structure model.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The patent segments the vessel structure into distinct segments within each graph representation, then compares and attaches corresponding segments across different frames. This segmentation allows for precise identification of consistent anatomical structures while excluding inconsistent portions caused by motion artifacts, improving overall representation reliability.

Inventive Principle:

Principle #1Segmentation

2Measurement precision

If information is aggregated from multiple frames to improve anatomical understanding, then diagnostic accuracy is improved, but inconsistencies reduce reliability of downstream tasks

Engineering Contradiction:

Improveaccuracy of anatomical representationVSAvoidreliability of downstream task results

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent implements a feedback mechanism where graph representations are processed in iterations, with each iteration using the results from previous iterations to refine the aggregated representation. The system compares segments across frames, identifies inconsistencies, and adjusts the aggregation process accordingly, ensuring that only consistent and reliable information is included in the final representation for downstream tasks.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent performs preliminary processing of individual frame graph representations before aggregation, including segment identification, spatial coordinate normalization, and consistency checking. This preliminary action ensures that the data is prepared and validated in advance, reducing inconsistencies that would otherwise affect downstream task reliability.

Inventive Principle:

Principle #10Preliminary action

3Adaptability or versatility

If graph representations are generated from consecutive medical images, then temporal information is captured, but inconsistencies arise due to varying imaging conditions

Engineering Contradiction:

Improvetemporal resolution of vessel structureVSAvoidprecision of vessel structure representation

Core Design Contradiction:

Adaptability or versatilityVSManufacturing precision

Solution Approach 1:

The patent dynamically adjusts the aggregation process to account for varying imaging conditions across frames. By comparing segments temporally and using spatial coordinates to match corresponding structures across different imaging angles and conditions, the system maintains precision while capturing temporal dynamics of the vessel structure.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS20250378558A1Aggregated representation of a vessel structure

Publication Date: 2025.12.11 SIEMENS HEALTHINEERS AG

AI summary

Graph representations of a vessel structure are received, each corresponding to a medical image and including a plurality of nodes and respective spatial coordinates for each node. The plurality of nodes are arranged according to at least one segment. For each respective segment of a first graph representation, it is determined whether a second graph representation of the plurality of graph representations includes a segment corresponding to the respective segment of the first graph representation and, if it is found that this is not the case, the second graph representation is augmented by adding an additional segment corresponding to the respective segment of the first graph representation. An aggregated representation of the vessel structure is generated depending on the augmented second graph representation.