Volumetric Filtering of Fluoroscopic Sweep Video for Device Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D volume reconstructions from CT scans of patient lungs are inaccurate due to deformation during procedures, causing challenges in guiding medical devices to targets accurately.
Innovation Solution
A method involving fluoroscopic imaging to generate a 3D volumetric reconstruction, filter and refine object positions, and interpolate gaps for precise localization of medical devices within the body.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If 3D volume reconstruction is generated from previously acquired CT scans, then navigation planning can be performed, but the accuracy deteriorates due to lung deformation during the procedure
Solution Approach 1:
The system performs preliminary action by acquiring fluoroscopic images at the beginning of the procedure to establish an initial 3D reconstruction, then continuously updates this reconstruction during the procedure using new fluoroscopic images. This allows the navigation system to maintain accuracy despite lung deformation over time, resolving the contradiction between having preliminary navigation data and maintaining accuracy during the procedure.
2Measurement precision
If fluoroscopic images are processed to generate 3D volumetric reconstruction, then real-time localization is achieved, but computational complexity and processing time increase
Solution Approach 1:
The system segments the fluoroscopic image processing into distinct stages: initial 3D reconstruction from a set of fluoroscopic images, then incremental updates using new images. This segmentation allows complex processing to be distributed over time, maintaining precision while managing computational complexity through phased processing rather than continuous full-reconstruction.
Solution Approach 2:
Instead of processing all fluoroscopic images continuously for full 3D reconstruction, the system performs partial action by using only the necessary subset of images for the current navigation step. This reduces computational complexity while maintaining sufficient localization precision for the immediate navigation task.
3Area of stationary object
If the entire 3D volume is processed for object detection, then complete coverage is achieved, but processing time and computational resources increase
Solution Approach 1:
The system applies local quality by focusing detection and processing resources on the specific region of interest where the medical device and target are located, rather than uniformly processing the entire 3D volume. This selective processing maintains detection coverage for the critical area while significantly improving processing speed and productivity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the accuracy of navigating medical devices to targets by providing real-time, precise localization and reducing false positives in fluoroscopic images.
Implementation Method 1
receiving a fluoroscopic video including a plurality of fluoroscopic images captured by a fluoroscope
Implementation Method 2
generating a three-dimensional (3D) volumetric reconstruction from the fluoroscopic images
Data Source
AI summary
Object detection including receiving a fluoroscopic video including a plurality of fluoroscopic images captured by a fluoroscope rotated about a patient, generating a three-dimensional (3D) volumetric reconstruction from the fluoroscopic images, deleting values in the 3D volumetric reconstruction beyond a region of interest about an object, generating 2D images from a remaining 3D volumetric reconstruction following the deleting of values: detecting the object in the 2D images to determine an initial position of the object, and refining the detected initial position of the object.