AI Segmentation for Vascular Imaging Without Mask Phase

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

X-ray based imaging methods, such as digital subtraction angiography, require two distinct phases for vascular imaging, leading to increased time and X-ray dose for the examination object, especially in 3D DSA.

Innovation Solution

A computer-implemented method that generates a vascular image data record by receiving a series of projection-X-ray images recorded in temporal succession, determining change image data records based on regions of interest, and generating the vascular image data record without the need for a mask phase, using time-intensity curves and threshold values to classify image points and reduce unnecessary X-ray exposure.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If digital subtraction angiography is used for vascular imaging, then vascular structures can be visualized, but time expenditure and X-ray dose increase due to requiring two distinct phases

Engineering Contradiction:

Improvevascular imaging accuracyVSAvoidexamination time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts and eliminates the mask phase from the traditional two-phase DSA process. By using only the filling phase images and applying AI-based segmentation to identify and remove non-vascular structures, the method retains vascular imaging capability while removing the time-consuming mask phase, thus resolving the contradiction between imaging accuracy and examination time

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent replaces the mechanical/subtractive image processing approach of traditional DSA with an AI-based segmentation system. The trained neural network automatically identifies vascular structures and separates them from non-vascular components, substituting the traditional mask-based subtraction method with a more efficient machine learning approach that reduces time expenditure

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If digital subtraction angiography is used for vascular imaging, then vascular structures can be visualized, but X-ray dose increases due to requiring two distinct phases

Engineering Contradiction:

Improvevascular imaging accuracyVSAvoidX-ray dose

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent extracts and eliminates the mask phase from the traditional two-phase DSA process. By using only the filling phase images and applying AI-based segmentation to identify and remove non-vascular structures, the method retains vascular imaging capability while removing the unnecessary mask phase exposure, thus reducing the total X-ray dose to the examination object

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent uses AI-based segmentation to create a digital representation of vascular structures from filling phase images alone. The trained neural network generates accurate vascular models without requiring physical mask images, effectively copying the essential vascular information while eliminating the need for mask phase X-ray exposure

Inventive Principle:

Principle #26Copying

3Loss of information

If 3D DSA is performed, then three-dimensional vascular information can be obtained, but time and X-ray dose expenditure increase significantly

Engineering Contradiction:

Improvethree-dimensional vascular informationVSAvoidexamination time

Core Design Contradiction:

Loss of informationVSLoss of time

Solution Approach 1:

The patent replaces the time-consuming multi-phase 3D DSA acquisition process with an AI-based segmentation approach applied to filling phase images. The trained neural network automatically extracts three-dimensional vascular information from the filling phase data, substituting the traditional multi-phase mechanical acquisition with an efficient computational method that maintains information quality while reducing time expenditure

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

4Loss of information

If 3D DSA is performed, then three-dimensional vascular information can be obtained, but X-ray dose increases significantly

Engineering Contradiction:

Improvethree-dimensional vascular informationVSAvoidX-ray dose

Core Design Contradiction:

Loss of informationVSObject-affected harmful factors

Solution Approach 1:

The patent extracts and eliminates the mask phase from the 3D DSA process. By using only the filling phase images and applying AI-based segmentation to obtain three-dimensional vascular information, the method maintains comprehensive vascular data while removing the unnecessary mask phase exposure, thus significantly reducing the total X-ray dose to the examination object

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS11823387B2Providing a vascular image data record

Publication Date: 2023.11.21 SIEMENS HEALTHINEERS AG

AI summary

A computer-implemented method for providing a vascular image data record includes receiving a plurality of projection-X-ray images recorded in temporal succession. The plurality of projection-X-ray images at least partially map a common examination region of an examination object. The plurality of projection-X-ray images map a temporal change in the examination region of the examination object. A change image data record is determined in each case based on at least one region of interest of the plurality of projection-X-ray images. The at least one region of interest includes a plurality of image points. The change image data record in each case includes a time-intensity curve for each of the image points. The vascular image data record is generated based on the change image data record, and the vascular image data record is provided.