Markerless Radiotherapy Calibration via 3D Point Cloud Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radiotherapy systems rely on marker placement for patient positioning, which is problematic in terms of repeatability and accuracy, especially for breathing measurement regions.
Innovation Solution
A markerless calibration and measurement system using a camera to compute a 3D point cloud representation of a patient, performing optimized principal components analysis, and determining measurement regions to establish a center point and other regions along the patient's torso, allowing for real-time positioning and compensation for parallax errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If marker placement systems are used for patient positioning, then positioning can be achieved, but repeatability and accuracy deteriorate
Solution Approach 1:
The patent replaces the mechanical marker placement system with an optical 3D imaging system using cameras to capture depth information. This substitution eliminates the need for physical markers while achieving comparable or superior positioning accuracy through computational methods, directly addressing the repeatability and accuracy issues of marker-based systems
Solution Approach 2:
The patent creates a digital 3D copy (point cloud representation) of the patient's anatomy from multiple camera views. This virtual model serves as the basis for all positioning measurements, eliminating the need for physical markers and improving repeatability through consistent digital replication of anatomical features
2Ease of operation
If marker-based positioning is used, then positioning can be performed, but device complexity increases due to marker placement requirements
Solution Approach 1:
The patent extracts and removes the marker components from the positioning system entirely. By using unmarked 3D imaging, the system eliminates the complexity of marker placement, removal, and repositioning while maintaining positioning capabilities through automated computational algorithms
Solution Approach 2:
The system automatically identifies anatomical features and establishes measurement regions through computational analysis of the 3D point cloud data. This self-service capability eliminates the need for manual marker placement and system setup, significantly simplifying the overall operation while maintaining precision
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
Enables accurate and repeatable patient positioning without markers, robust to patient movement and clutter, and sensitive to physiological constraints for enhanced data uniformity and orientation calculation.
Implementation Method 1
a camera to obtaining a plurality of depth points (X, Y, z) of the patient, where the camera includes pixel coordinates (x, y) and depth to target z
Implementation Method 2
computing a point cloud covariance and a point cloud mean, where the covariance and the mean of the point cloud are used to perform an optimized principle components analysis
Implementation Method 3
compensating for a parallax error in real-time during a calibration measurement or during a radiotherapy treatment plan
Data Source
AI summary
A method of real-time markerless measurement and calibration of a patient positioning system includes computing a three dimensional point cloud representation of a patient by using a camera to obtain depth points of the patient, where the z-direction corresponds to a direction along the focal length of the camera, demanding a z-limit that corresponds to a lower (z-min) and an upper (z-max) specification, computing a point cloud covariance and a point cloud mean for an optimized principle components analysis to compute a torso topography of the patient, determining a center point of the point cloud as an anchor measurement region and using the camera to determine a distance to the patient to establish an angle between a patient torso and a vector that is perpendicular to the camera, averaging the depth measurements in reference regions and compensating for parallax error in real-time during a calibration measurement or radiotherapy treatment plan.