3D Image Registration Using Depth Camera and Positioning Tool
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image-guided surgical navigation systems require manual registration of feature points, which increases surgery preparation time, reduces accuracy, and poses a risk of cross-infection due to direct contact with the patient.
Innovation Solution
A device and method for automatic 3D image registration using a depth recognition camera and positioning device, which acquires 3D depth data and image data to perform surface registration, extracting camera reference coordinates of a landmark and converting relative position information to align the patient's image and positioning device coordinate systems without direct skin contact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual registration method is used to acquire landmark, then registration can be performed, but preparation time is increased and accuracy is reduced
Solution Approach 1:
The system performs automatic self-registration by capturing images of the registration body, automatically extracting feature points through image processing, and computing transformation parameters without requiring manual intervention. This eliminates the dependency on technician skill level and reduces preparation time while maintaining high registration accuracy.
2Measurement precision
If contact manner scanning is used to register physical structure, then registration can be performed, but cross-infection risk occurs
Solution Approach 1:
A registration body with a registration marker is introduced as an intermediary object. The registration marker is captured by the depth recognition camera without direct contact with the patient's body, yet it enables accurate registration by serving as a mediator between the imaging system and the patient anatomy. This eliminates cross-infection risk while maintaining registration accuracy.
3Measurement precision
If manual orientation of feature points is performed, then registration can be achieved, but it depends on technician skill level
Solution Approach 1:
The system automatically captures images of the registration body, processes the images to extract feature points, and computes transformation parameters without requiring manual orientation by technicians. This eliminates the dependency on technician skill level and simplifies the operation while maintaining high registration accuracy through automated image processing algorithms.
Data Source
AI summary
A device for providing 3D image registration includes: a collection unit acquiring 3D depth data of a part and an image of a registration body from a depth recognition camera and acquiring three-dimensional coordinates of a positioning tool on the registration body from a positioning device; a first registration unit performing surface registration of a pre-stored 3D medical image of patient and the 3D depth data; a second registration unit extracting camera reference three-dimensional coordinates of a landmark attached to the registration body from the image of the registration body and converting pre-stored relative position information of the landmark with reference to the three-dimensional coordinates of the positioning tool, to perform registration of camera reference three-dimensional coordinates of the landmark and the converted position information thereof; and a third registration unit performing final registration by using results of registration performed by the first registration unit and the second registration unit.