Multi-Camera Optical Tracking for MRI Motion Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical imaging technologies, particularly MRI, face challenges in accurately tracking motion across a wide range of positions, leading to motion artifacts that degrade image quality, especially in neuroimaging where patients, especially pediatric and stroke patients, often move, necessitating repeat scans and reduced diagnostic confidence.
Innovation Solution
The use of multiple cameras integrated into the MRI scanner, with a self-encoding marker and synchronized video data processing, allows for expanded tracking range and accurate pose determination, enabling real-time motion correction by calculating the object's pose from partial views and combining data using algorithms like the augmented DLT, ensuring robust motion tracking across various subjects and motion ranges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single camera and single marker are used for optical tracking, then the system is simple and easy to operate, but the tracking range is limited and cannot accommodate subjects with varying head sizes and motion ranges
Solution Approach 1:
The system divides the tracking task into multiple independent camera units, each responsible for a specific field of view. Multiple cameras (at least two) are positioned to capture different regions, allowing the system to track objects across a wider range of positions while maintaining manageable complexity through modular architecture
Solution Approach 2:
The marker design incorporates self-encoding features that enable it to be recognized and tracked from multiple camera angles simultaneously. The marker can encode its position and orientation information in a way that is universally readable by any camera in the system, allowing a single marker to serve multiple tracking functions across different viewing angles
2Measurement precision
If the marker must be fully visible for accurate tracking, then tracking precision is maintained, but the system fails when line of sight is obscured by head coils or patient hair
Solution Approach 1:
The self-encoding marker is designed to provide sufficient pose determination information from partial views rather than requiring complete visibility. The marker encodes enough geometric and identification data in each visible portion that accurate tracking can be achieved even when parts of the marker are obscured by head coils, patient hair, or other obstacles
Solution Approach 2:
The marker incorporates distinctive visual patterns, colors, or encoding schemes that enable reliable identification and pose calculation from partial observations. These visual features allow the tracking system to distinguish the marker from surrounding elements and accurately determine its position and orientation even when the full marker is not visible
3Speed
If cameras are placed close to the patient's head to improve tracking, then temporal resolution is high, but the field of view is limited and part of the marker lies outside the camera view
Solution Approach 1:
The system transitions from a single-camera perspective to a multi-camera spatial arrangement, effectively adding dimensional coverage. By positioning multiple cameras at different locations around the imaging area, the system achieves comprehensive spatial coverage without requiring any single camera to be placed in suboptimal positions, thus maintaining high temporal resolution while expanding the effective field of view
Data Source
AI summary
Methods to quantify motion of a human or animal subject during a magnetic resonance imaging (MRI) exam are provided. In particular, these algorithms make it possible to track head motion over an extended range by processing data obtained from multiple cameras. These methods make current motion tracking methods more applicable to a wider patient population.