3D Flow Map Reconstruction from Rotational Angiography
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional 2D methods for deriving flow information from angiographic images fail in the presence of overlapping vasculature and implants, especially when using optical flow approaches with rotational acquisitions, which are limited by angular differences between projection frames.
Innovation Solution
A framework that acquires and reconstructs four-dimensional (4D) DSA data to generate three-dimensional (3D) flow maps by registering temporally adjacent projection images, determining 2D flow vectors, sorting them by heart phases, and reconstructing 3D flow maps with flow vectors representing magnitude and direction (x′, y′, z′) to account for angular differences and provide quantitative metrics on blood/contrast mixture flow direction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D optical flow approach is used to derive flow vectors from angiographic images, then flow magnitude and direction can be determined, but the method fails in the presence of overlapping vasculature and implants
Solution Approach 1:
The patent transitions from 2D flow analysis to 3D flow map reconstruction by incorporating rotational acquisition data. Multiple projection images acquired at different angles are combined to reconstruct three-dimensional flow vectors, allowing visualization of flow direction in 3D space and resolving ambiguities caused by overlapping vasculature in single 2D projections
Solution Approach 2:
The patent segments the vascular structure into multiple angular projections, acquiring images at different rotational positions. Each projection provides flow information from a specific angle, and these segmented views are then integrated to form a complete 3D flow map, avoiding the limitations of any single 2D view
2Area of stationary object
If rotational acquisitions are used to improve vascular visualization, then more complete vascular structures can be captured, but angular differences between projection frames limit the accuracy of optical flow approaches
Solution Approach 1:
The patent resolves the angular difference problem by reconstructing 3D flow maps from multiple angular projections. Instead of attempting to calculate optical flow between frames at different angles (which is inaccurate), the system reconstructs the flow field in three-dimensional space, where vectors from different projections can be accurately integrated to compensate for angular variations
Solution Approach 2:
The patent introduces an intermediate 3D flow map reconstruction step that mediates between the angularly-differentiated projection images. This intermediate representation serves as a common coordinate system where flow vectors from different angles can be accurately combined, eliminating the direct optical flow calculation between angularly-separated frames
3Productivity
If 2D flow maps are used to represent blood flow, then flow information can be visualized, but overlapping vasculature causes interpretation difficulties
Solution Approach 1:
The patent resolves the interpretation difficulty of overlapping vasculature by transitioning from 2D to 3D flow map visualization. The three-dimensional representation allows observers to distinguish between vessels at different spatial locations and depths, making it possible to accurately interpret flow direction even when vessels overlap in the projection view
Data Source
AI summary
Described herein are technologies for facilitating reconstruction of flow data. In accordance with one aspect, the framework receives a four-dimensional projection image dataset and registers one or more pairs of temporally adjacent projection images in the image dataset. Two-dimensional flow maps may be determined based on the registered pairs. The framework may then sort the two-dimensional flow maps according to heart phases, and reconstruct a three-dimensional flow map based on the sorted two-dimensional flow maps.