Automatic Device Unlocking Detection in Fluoroscopic Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for navigating and positioning guidewires and interventional devices in vascular procedures are tedious, time-consuming, and require additional contrast agent bursts, with challenges in tracking anatomical landmarks and device motion due to scatter in fluoroscopic images, leading to inaccuracies and increased exposure for patients and medical staff.
Innovation Solution
A method for automatic device unlocking detection using anatomical landmarks coupled with device-based motion compensation, which involves receiving a sequence of fluoroscopic images, detecting a device, determining its motion field, generating integrated images, calculating a saliency metric, and identifying landmarks based on this metric to assess movement relative to the device, thereby enabling fully automatic dynamic motion compensation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If anatomical landmarks are tracked for motion compensation, then navigation accuracy is improved, but the task becomes very difficult due to faint visibility and clutter from opaque objects
Solution Approach 1:
The patent uses an interventional device as an intermediary object to track motion. Instead of directly tracking faint anatomical landmarks, the system tracks the clearly visible device which has a known spatial relationship to the landmarks, thereby indirectly obtaining landmark motion information while avoiding the difficulty of direct landmark detection
Solution Approach 2:
The system creates a virtual model (copy) of the anatomical structure based on pre-acquired 3D imaging data. This virtual model is then registered and tracked alongside the physical device, allowing motion compensation without requiring direct detection of faint anatomical features in real-time fluoroscopic images
2Measurement precision
If device motion is tracked for motion compensation, then navigation support is improved, but device unlocking must be detected to assess validity
Solution Approach 1:
The system continuously monitors the spatial relationship between the tracked device and the virtual anatomical model, providing feedback on whether the device remains in stable contact with the anatomy. When device unlocking is detected through changes in this feedback signal, the system can assess the validity of motion compensation and trigger appropriate responses
Solution Approach 2:
The system automatically detects device unlocking events and assesses motion compensation validity without requiring additional user input or manual verification. The motion tracking data itself provides the information needed to determine whether the device has moved from its intended position
3Device complexity
If static motion compensation is used with a single geometric transformation, then processing simplicity is improved, but navigation accuracy deteriorates due to anatomical motion during intervention
Solution Approach 1:
The patent implements dynamic motion compensation by continuously tracking the device through multiple fluoroscopic frames and updating the geometric transformation for each frame. This dynamic approach adapts to anatomical motion during the intervention, maintaining navigation accuracy whereas static compensation would fail to account for physiological movements
Data Source
AI summary
A method and device is proposed for automatic detection of an event in which a device is leaving a stable position relative to and within an anatomy. The method comprises the steps of receiving a sequence of fluoroscopic images, detecting a device in at least two of the fluoroscopic images, determining a motion field of the detected device in the sequence of fluoroscopic images, generating a sequence of integrated images by integrating the sequence of fluoroscopic images taking into consideration the motion field, determining a saliency metric based on the integrated images, identifying a landmark in the integrated images based on the saliency metric, and determining as to whether the landmark is moving relative to the device, based on a variation of the saliency metric.