AI Pose Correction for Medical Imaging Patient Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Patient positioning in medical imaging and therapy workflows is highly manual, subjective, and error-prone, relying heavily on technician experience, leading to sub-optimal image quality and ineffective therapies.
Innovation Solution
A computer-implemented method using artificial neural networks to determine a target patient pose from a positioning goal, correct the current pose, and provide feedback through textual instructions or animations, leveraging trained medical and anatomical knowledge to automate and improve positioning accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual patient positioning is performed by technicians, then positioning can be adjusted based on experience, but the process is time-consuming and error-prone
Solution Approach 1:
The patent replaces the manual mechanical positioning process with an automated computer vision system using neural networks. The system automatically detects patient pose from images, compares it to target poses, and provides correction guidance, eliminating the need for manual assessment and reducing both time and errors.
Solution Approach 2:
The system enables self-positioning by providing automated feedback to patients through visual and textual guidance. Patients can independently adjust their pose based on real-time feedback from the system, reducing dependency on technician intervention and accelerating the positioning process.
2Extent of automation
If automated pose correction is implemented using neural networks, then positioning accuracy and automation are improved, but system complexity increases
Solution Approach 1:
The neural network system is designed to handle multiple positioning scenarios and body parts through a single unified model. The system can process different types of medical imaging modalities and provide pose correction for various anatomical regions, reducing the need for multiple specialized systems.
Solution Approach 2:
The patent introduces an intermediary layer that translates complex neural network outputs into simple, actionable visual and textual feedback for patients. This intermediary interface simplifies the user experience while maintaining the sophisticated automation capabilities of the underlying system.
3Measurement precision
If continuous feedback and iterative correction are provided, then positioning precision is improved, but processing time and computational load increase
Solution Approach 1:
The system implements iterative correction cycles that continue only until a predefined accuracy threshold is met or a maximum number of iterations is reached. This prevents excessive computational effort while ensuring sufficient positioning precision for clinical requirements.
Solution Approach 2:
The neural network is pre-trained on large datasets of correct poses during an offline training phase. This preliminary action transfers learned knowledge to the deployment phase, where the system only needs to perform lightweight inference and comparison operations, significantly reducing real-time computational energy requirements.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for correcting a current pose (28) of a patient (26) for a diagnostic image capture procedure, comprising: receiving (104) an input image of the patient (26); determining (102), by means of a first artificial neural network, a target patient pose (22) from a positioning goal (18); determining (106) a current pose (28) of the patient from the input image and determining (108) a correction plan (30) for correcting the pose of the patient (26) from the current patient pose (28) and the target patient pose (22). Furthermore, the invention relates to a device, a computer program product and a computer-readable medium comprising instructions for carrying out the method.