Anatomical Structure Registration by Filtering Surgical Image Noise
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing registration methods for preoperative models in surgical planning lack accuracy due to the inclusion of soft-tissue and artificial structures during surgery, which hinder precise alignment with the surgical field.
Innovation Solution
A device and method for iteratively filtering preoperative model data points based on distance thresholds to align with 3D surgical images, excluding non-target structures and enhancing registration accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the preoperative model includes all data points from medical images, then the model is complete, but soft-tissue and artificial structures create noise that reduces registration accuracy
Solution Approach 1:
The patent extracts and removes harmful data points representing soft-tissue and artificial structures from the preoperative model. The system identifies and discards these unwanted elements while retaining only the relevant anatomical structure data points, thereby eliminating noise that degrades registration accuracy without losing essential anatomical information.
Solution Approach 2:
The patent applies different quality criteria to different regions of the preoperative model. Rather than uniformly treating all data points, the system selectively filters data points based on their local characteristics and relevance to the target anatomical structure. This allows the model to maintain high quality in critical regions while removing noisy data from non-relevant areas.
2Measurement precision
If fiducial markers are implanted for registration, then registration accuracy is improved, but surgical invasiveness increases
Solution Approach 1:
The patent extracts registration capability directly from the anatomical structure data itself, eliminating the need for separate fiducial markers. By deriving registration features from the preoperative model and intraoperative images of the anatomical structure, the system removes the requirement for invasive marker implantation while maintaining registration functionality.
Solution Approach 2:
The patent uses image-based copying to create a digital representation of the anatomical structure for registration purposes. Instead of physically implanting markers, the system captures images of the exposed anatomical structure during surgery and uses these image copies for alignment and registration, replacing invasive physical markers with non-invasive optical copies.
3Productivity
If the preoperative model is used directly without filtering, then the registration process is faster, but the alignment with surgical field is inaccurate
Solution Approach 1:
The patent performs preliminary filtering and identification of relevant data points in the preoperative model before the actual registration process. By pre-processing the model to identify and retain only relevant anatomical data points while removing soft-tissue and artificial structures, the system prepares an optimized dataset that enables both fast and accurate registration without requiring extensive filtering during the surgical procedure.
Data Source
AI summary
A method and a device for segmentation of a preoperative model of a target anatomical structure in view of the registration of said preoperative model with an image of a target anatomical structure exposed during surgery.