3D Vessel Model Generation Using Probability Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating 3D vessel models from 2D angiograms are prone to inaccuracies due to user-dependent input, leading to time-consuming and fatiguing interaction processes that may result in misinterpretation and loss of concentration.
Innovation Solution
A method that utilizes probability maps to enhance the accuracy of 2D centrelines and bifurcation point identification by providing users with a visual representation of feature likelihood, allowing for more precise placement of points of interest and automatic snapping to detected positions, thereby reducing user effort and improving 3D model quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual segmentation and point indication is used to generate 3D vessel models, then the accuracy of 2D centrelines can be improved, but the interaction time and user fatigue increase significantly
Solution Approach 1:
The system performs preliminary automatic segmentation and generates preliminary 2D centrelines before user interaction. This preliminary action provides a head start, reducing the amount of manual work required while maintaining high accuracy through subsequent user refinement of the pre-computed results.
Solution Approach 2:
The system enables users to indicate points on the probability map rather than directly on the angiogram. This self-service approach leverages the probability map's guidance to automatically improve point accuracy without requiring users to have expert-level visual interpretation skills, thus reducing interaction time while maintaining precision.
2Reliability
If multiple 2D angiograms are manually segmented to construct 3D vessel models, then the reliability of the 3D model can be improved, but the complexity of the operation increases
Solution Approach 1:
The probability map serves as an intermediary between the raw angiogram and the user's point indication. It translates complex image data into simplified visual guidance, making the segmentation operation easier while maintaining reliability through the automated computation of feature probabilities that guide accurate point placement.
Solution Approach 2:
The system changes the parameter space by transforming the raw pixel intensity data into probability values that represent the likelihood of vessel features. This parameter transformation simplifies the user's task by providing intuitive visual cues about where to place points, reducing operational complexity while improving reliability.
3Measurement precision
If users indicate points directly on angiograms to extract centrelines, then the precision of point location can be improved, but the difficulty of detecting and measuring increases due to image complexity
Solution Approach 1:
The probability map acts as an intermediary layer that simplifies the detection task. Instead of directly analyzing complex angiogram images with overlapping vessels and low contrast, users interact with the probability map that has already processed and highlighted likely feature locations, reducing detection difficulty while maintaining precision.
Solution Approach 2:
The system replaces the mechanical process of visual inspection and manual point placement on complex images with an automated probability computation system. This substitution uses algorithmic processing to pre-identify likely feature locations, reducing the cognitive and visual burden on users while maintaining high measurement precision.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The invention relates to the generation of a 3D vessel model of a region of interest of an object. The invention relates particularly to a medical imaging system and a method for generating a 3D vessel model of a region of interest of an object and to a computer program element as well as a computer readable medium for generating a 3D vessel model of a region of interest of an object. In order to facilitate and improve the generation of a 3D vessel model of a region of interest of an object, a medical imaging system and a method with the following steps is provided: Acquiring at least two 2D X-ray projection images of contrast enhanced vascular structures from different viewing angles; determining a probability map for predetermined vessel features for each of the 2D X-ray projection images; displaying the probability map for each of the 2D X-ray projection images for interaction purposes; segmenting vessels of interest by indicating the location of a first set of points of interest in the probability map of one of the at least two 2D X-ray projection images by interaction of the user, determining and displaying epipolar lines for the first set of points of interest in the probability map of the other one of the at least two 2D X-ray projection images, indicating the location of a second set of points of interest in the probability map of the other one of the at least two 2D X-ray projection images by the user, wherein the epipolar lines act as orientation and wherein the second set of points is corresponding to the indicated first points, determining the closest relevant predetermined features of the vascular structure upon the indication of the location of points of interest and extracting2D centrelines,bifurcation points and/or vessel borders of the vascular structure from the determined features of the vascular structure;and calculating a 3D vessel model from the extracted 2D vessel centrelines,bifurcation points and/or vessel borders.