Anatomical Structure Registration Using Iterative Surface Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for registering a preoperative model of a target anatomical structure during surgery are not optimal, particularly in minimally invasive procedures like robotic-assisted orthopedic surgery, where precise alignment between the preoperative and intraoperative frames is crucial.
Innovation Solution
A device and method utilizing iterative registration of selected sub-sets of points from a preoperative model onto 3D images of the surgical field, filtering out points corresponding to surrounding tissues or artificial structures, and using a predefined threshold to achieve optimal alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional registration methods using fiducial markers are used, then registration accuracy is achieved, but surgical invasiveness increases
Solution Approach 1:
The invention extracts and removes the fiducial markers from the registration process. Instead of using external markers attached to the bone, the system directly uses the anatomical surface of the bone itself as the registration target, thereby eliminating the invasive marker implantation step while maintaining registration accuracy through direct surface matching
Solution Approach 2:
The invention creates a digital copy (3D model) of the bone's anatomical surface from preoperative imaging data and matches it with the actual intraoperative bone surface. This digital replica serves as the registration reference, replacing the need for physical fiducial markers while preserving measurement precision through surface geometry correspondence
2Area of stationary object
If all data points from the preoperative model are used for registration, then comprehensive coverage is achieved, but registration precision decreases due to inclusion of surrounding tissues and artificial structures
Solution Approach 1:
The invention segments the set of data points from the preoperative model into two distinct groups: points corresponding to the target anatomical structure (bone surface) and points corresponding to surrounding tissues or artificial structures. This segmentation allows selective use of only the relevant bone surface points for registration, improving precision while maintaining adequate coverage of the anatomical target
Solution Approach 2:
The invention applies different quality criteria to different regions of the preoperative model. Specifically, it identifies and selects only those data points that correspond to the exposed bone surface with appropriate geometric properties, while excluding points from soft tissues, implants, or other non-bone structures. This local quality filtering ensures high registration precision by using only the most suitable data points
3Measurement precision
If iterative registration with multiple iterations is performed, then registration accuracy is improved, but computational time increases
Solution Approach 1:
The invention performs preliminary actions before the main iterative registration process: it pre-segments and filters the data points from the preoperative model to identify only those corresponding to the target bone surface, and establishes an initial coordinate system alignment. These preliminary steps reduce the complexity of the subsequent iterative matching by providing a refined, pre-processed dataset and a better starting point for iteration
Solution Approach 2:
The invention applies partial action by using a subset of data points (only those corresponding to the bone surface) rather than all available points from the preoperative model. This selective use of data reduces the computational burden of each iteration while maintaining registration accuracy, as the relevant bone surface points provide sufficient information for precise alignment without the noise from unrelated structures
Data Source
Figure 1
Figure 2~3
Figure 4(a)~4(c)
AI summary
The present invention relates to a method and a device for segmentation of an exposed target anatomical structure in view of an iterative registration during surgery. Registration is performed by iteratively expanding the data used for registration along the longitudinal axis of a preoperative model of said target anatomical structure.