AR Fracture Reduction Path Planning With 3D RL Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional fracture reduction surgery relies heavily on a doctor's experience and two-dimensional imaging, leading to prolonged operation times, increased surgical risk, and difficulty in achieving accurate fracture alignment, especially for complex fractures.
Innovation Solution
A method and system using augmented reality technology to reconstruct a three-dimensional model of the fracture site, combined with a cascaded reinforcement learning module and multi-physics field simulation optimization, to plan and optimize a fracture reduction path, ensuring accurate and safe alignment of bone blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional two-dimensional image-assisted technology is used for fracture reduction, then the surgical operation can be performed with existing equipment, but the surgical accuracy is insufficient and operation time is prolonged
Solution Approach 1:
The patent transitions from two-dimensional imaging (X-rays) to three-dimensional visualization by reconstructing a 3D model of the fracture site from CT scan data. This dimensional change allows surgeons to view the fracture structure from multiple angles and understand spatial relationships between bone fragments, thereby improving reduction accuracy and reducing the need for repeated trial adjustments during surgery.
Solution Approach 2:
The system performs preliminary planning by pre-reconstructing the 3D fracture model and pre-calculating the optimal reduction path before the actual surgery. The reinforcement learning module generates a planned reduction trajectory based on the pre-processed 3D model, allowing the surgeon to follow a predetermined optimal path during the operation, thus improving accuracy and reducing operation time.
2Reliability
If doctors rely on visual feedback and experience for fracture reduction, then the surgical procedure can be performed with simple equipment, but the surgical risk increases and accuracy cannot be ensured
Solution Approach 1:
The patent introduces a computational intermediary system that bridges the gap between 3D imaging data and surgical execution. The reinforcement learning module acts as an intelligent mediator that processes the 3D fracture model and generates an optimized reduction path, providing objective guidance that reduces reliance on subjective visual feedback and doctor experience, thereby improving surgical safety.
Solution Approach 2:
The system replaces the mechanical/visual feedback mechanism with a computational simulation and planning mechanism. Instead of relying on real-time visual assessment and tactile feedback during surgery, the system uses pre-computed 3D models and reinforcement learning algorithms to determine the optimal reduction path, reducing the impact of human factors and improving reliability.
3Manufacturing precision
If repeated attempts and adjustments are made during fracture reduction, then the fracture can be reduced, but the operation time is prolonged and surgical risk increases
Solution Approach 1:
The reinforcement learning module incorporates feedback mechanisms where the planned reduction path is continuously evaluated against the 3D fracture model. The system uses reward functions that evaluate alignment accuracy at each step, allowing the planning algorithm to automatically adjust and optimize the reduction trajectory, ensuring high precision while minimizing the number of trial adjustments needed during surgery.
Data Source
AI summary
A method for auxiliarily planning a fracture reduction path based on an augmented reality technology includes: obtaining fracture image data for a fracture area of a patient, the fracture image data including computer tomography (CT) data and X-ray data; reconstructing a three-dimensional model of a fracture part based on the fracture image data; matching and calibrating the three-dimensional model of the fracture part based on the fracture area of the patient in an actual scene; inputting a calibrated three-dimensional model of the fracture part and a corresponding fracture type and fracture surgery requirement into a fracture reduction path planning model to determine at least one corresponding optimized fracture reduction path, where the fracture reduction path planning model includes a cascaded reinforcement learning (RL) module and a multi-physics field simulation optimization module; and rendering each fracture reduction path, and determining a target fracture reduction path according to a detected interactive instruction.