Back to results
US 2019/0172211 A1Published application
Method for automatically detecting systemic arteries in arbitrary field-of-view computed tomography angiography (cta)
AssigneesAgfa Healthcare NV, \/R\/Is Zentrum FüR Virtual Reality Und Visualisierung
InventorsAlexey NOVIKOV, David MAJOR, Maria WIMMER, Katja BUEHLER
Life of this patent
Estimating the term…
Abstract
A method and system for automatically detecting systemic arteries in arbitrary field-of-view computed tomography angiography (CTA) includes fully-automatically analyzing a medical image represented by a digital image representation.
Claims9
1-9. (canceled)
10:
- A method for analyzing a slice-based scanned image volume of a body represented by a digital image representation, the method comprising:
- determining a body region to which the slice-based scanned image volume belongs;
- determining expected vessel characteristics for vessels that are located in the determined body region, the expected vessel characteristics including vessel radii, thresholds for a measure of circularity of the respective vessels at a given position, vessel intensity value thresholds, and anatomy-based vessel segment connection rules;
- determining seed points in the image volume representing centers of vessels;
- tracking unconnected vessel segments with the seed points as a starting point by iteratively determining subsequent neighboring seed candidates which are selected based on a value representing a likeliness for a point of a slice-based representation belonging to one of the vessels including the expected vessel characteristics;
- and
- iteratively extending a vessel tree by connecting the unconnected vessel segments, applying a set of connection rules to the unconnected vessel segments in each iteration of the iteratively extending step based on the expected vessel characteristics, and terminating the iteration when none of a set of connection rules is fulfilled.
4 more elements
11:
- The method according to claim 10, wherein the step of determining the body region to which the slice-based scanned image volume belongs includes using a slice-based regression method applied to tissue composition in slices of the slice-based scanned image volume.
12:
- The method according to claim 10, wherein the step of determining the seed points in the image volume representing centers of vessels includes:
- pre-selecting candidate seed points by excluding seed points having an intensity below a predetermined threshold;
- detecting a shape of vessel-shaped objects through edge detection of the vessels through gradient estimation;
- and
- seed filtering by excluding candidate vessel positions which do not have a sufficiently circular shape in a transversal plane, which is evaluated by determining whether each of the vessel positions has a relative distance parameter below a predetermined threshold.
2 more elements
13:
- The method according to claim 10, wherein the step of tracking the unconnected vessel segments is initiated by a user selecting one of the determined seed points.
14:
- The method according to claim 10, further comprising associating a position of slice ranges to the body region by using machine learning techniques.
15:
- The method according to claim 10, wherein seed points are excluded as the seed candidates based on a comparison of a diameter value associated with the respective seed candidate against expected diameter values for the body region where the seed points are determined in the step of determining the seed points.
16:
- The method according to claim 10, further comprising optimizing an extension of the vessel tree by connecting the unconnected vessel segments using a genetic algorithm to optimize body-part specific parameters that define the vessel segment connection rules.
17:
- A non-transitory computer readable medium including a computer program product adapted to carry out the method of claim 10 when run on a computer.