3D Voxel Marker Registration for CT-Guided AR Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing registration methods for augmented reality in surgery, using IR and ArUco markers, suffer from inaccuracies due to errors in central coordinates and invisibility in CT scans, leading to potential surgical risks.
Innovation Solution
A method and apparatus utilizing a 3D voxel marker with multiple densities for precise registration, involving CT scanning, segmentation, and two-stage registration to align the voxel marker with a template marker, minimizing brightness differences between voxels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2D ArUco markers are printed on marker surfaces, then the markers can be visually detected, but the markers cannot be detected in CT scans and patterns become invisible
Solution Approach 1:
The marker design transitions from 2D printed patterns to 3D voxel structures with varying material densities. The density parameter is changed to ensure CT scan visibility while maintaining the marker's detection capability through multiple density levels that create distinguishable patterns in volumetric scans.
Solution Approach 2:
The marker evolves from a 2D printed surface pattern to a 3D volumetric structure. By adding the third dimension and using varying densities within the voxel grid, the marker becomes visible in CT scans while preserving pattern recognition capabilities through spatial arrangement of different density materials.
2Productivity
If point-based registration using IR markers is performed, then registration can be completed, but errors in central coordinates occur due to CT slice interval and imaging direction
Solution Approach 1:
The registration process is divided into two stages: primary registration using rigid body transformation for quick alignment, and secondary registration minimizing brightness differences for precision. The voxel marker is also segmented into multiple density regions that provide multiple reference points for accurate coordinate determination.
Solution Approach 2:
The voxel marker is designed with predetermined density patterns and geometric features before scanning. This preliminary structuring allows the registration algorithm to quickly identify reference points and perform accurate coordinate mapping without being affected by CT slice interval variations.
3Device complexity
If a single-density voxel marker is used, then the marker structure is simple, but the marker cannot provide precise registration information
Solution Approach 1:
Different regions of the voxel marker are assigned different material densities based on their functional requirements. High-density materials are placed in regions requiring strong CT signal, while low-density materials are used in regions needing subtle differentiation. This local variation in density provides rich registration information while maintaining an overall simple cubic structure.
Solution Approach 2:
The voxel marker is constructed from multiple materials with different densities, creating a composite structure. This composite approach allows the marker to simultaneously provide strong CT visibility through high-density components and fine differentiation through low-density components, achieving high registration precision without excessive structural complexity.
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
Provides precise registration between scan and actual models, enhancing surgical accuracy by reducing registration errors and improving surgical safety.
Implementation Method 1
performing a CT scan of a target object to which a voxel marker is attached
Data Source
AI summary
The present disclosure relates to a method, apparatus, and computer program for registering an augmented reality object, based on a voxel marker. The method of registering an augmented reality object, based on a voxel marker includes performing a CT scan of a target object to which the voxel marker is attached, performing segmentation by which only a region of interest of the target object and the voxel marker are extracted from a CT scan image, reconstructing the segmented CT scan image into a 3D model, and performing registration by aligning a voxel marker generated by the CT scan to the location, size, and rotation of a template voxel marker.