Vascular Constriction Detection with Differential Images and Neural Networks

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for locating vascular constrictions in angiographic images are time-consuming and prone to missing small constrictions, potentially delaying vital interventions.

Innovation Solution

A computer-implemented method that involves computing differential images from a temporal sequence of angiographic images, identifying temporal sequences of sub-regions where contrast agent enters and leaves, and using a neural network to classify these sequences to locate vascular constrictions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a radiologist manually reviews angiographic images by zooming in and out to identify vascular constrictions, then the process allows for detailed visual inspection, but it is time-consuming and small constrictions may be overlooked

Engineering Contradiction:
Improvedetection accuracyVSAvoidtime to locate constriction
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent replaces the manual mechanical process of radiologist review (zooming in and out visually) with an automated computer-based system that processes angiographic images through differential computation and neural network classification, thereby eliminating manual time investment while maintaining or improving detection accuracy

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

Solution Approach 2:

The patent introduces differential images as an intermediary representation that highlights changes in contrast agent flow between consecutive angiographic frames. This intermediary form makes vascular constrictions more visually prominent and easier to detect automatically, bridging the gap between raw image data and accurate constriction localization

Inventive Principle:
Principle #24Intermediary (Mediator)

2Area of stationary object

If CT angiogram is used to identify vascular constriction location, then the scan provides comprehensive brain imaging, but the small size of vascular constrictions hampers determination

Engineering Contradiction:
Improveimaging coverageVSAvoidconstriction detection capability
Core Design Contradiction:
Area of stationary objectVSMeasurement precision

Solution Approach 1:

The patent transforms the imaging parameter from static anatomical visualization (CT angiogram) to dynamic functional flow visualization by computing differential images that capture temporal changes in contrast agent movement. This parameter change enables detection of small vascular constrictions that are invisible in standard anatomical scans

Inventive Principle:
Principle #35Parameter changes

3Reliability

If CT perfusion scan is performed to identify affected brain regions, then the scan indicates regions with insufficient blood flow, but it only identifies a section of the brain rather than a specific vascular branch

Engineering Contradiction:
Improveperfusion assessment accuracyVSAvoidvascular branch localization
Core Design Contradiction:
ReliabilityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent segments the vasculature into multiple sub-regions and processes each sub-region independently through differential image computation and neural network classification. This segmentation approach enables precise localization to specific vascular branches while maintaining reliable perfusion assessment at the regional level

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4268183B1Locating vascular constrictions
Publication Date: 2025.08.27 KONINKLIJKE PHILIPS NV
  • EP4268183B1 patent drawingFigure 1~3
  • EP4268183B1 patent drawingFigure 4
  • EP4268183B1 patent drawingFigure 5~6

AI summary

A computer-implemented method of locating a vascular constriction in a temporal sequence of angiographic images, includes identifying (S130), from a temporal sequence of differential images, temporal sequences of a subset of sub-regions (120i,j) of the vasculature wherein contrast agent enters the sub-region, and the contrast agent subsequently leaves the sub-region; and inputting (S140) the identified temporal sequences of the subset into a neural network (130) trained to classify, from temporal sequences of angiographic images of the vasculature, a sub-region (120i,j) of the vasculature as including a vascular constriction (140).