Ultrasound Probe Tracking From Image Frames for Exam Workflow Logs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current ultrasound imaging systems lack effective methods to accurately monitor and record sonographer performance during exams, limiting the ability to enhance user feedback and improve workflow efficiency.
Innovation Solution
An ultrasound imaging system that utilizes neural networks to determine probe positions and orientations based on image frames, logs user inputs, and generates comprehensive records of user performance, including narrative summaries for review by sonographers and radiologists.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If log files capture only user inputs (button pushes, keystrokes), then system workflow analysis is possible, but probe manipulation performance cannot be monitored
Solution Approach 1:
The patent introduces an intermediary system that captures probe manipulation data through the existing ultrasound imaging system's image frames. The neural network acts as a mediator that translates standard image data into probe position and orientation information without requiring separate sensors or complex hardware modifications.
Solution Approach 2:
The patent replaces potential mechanical tracking systems with a computational approach using neural networks. Instead of using mechanical sensors or trackers on the probe, the system uses image processing and AI algorithms to infer probe manipulations from ultrasound images, reducing hardware complexity.
2Measurement precision
If comprehensive probe manipulation tracking is implemented, then sonographer performance monitoring is improved, but data processing complexity increases
Solution Approach 1:
The system uses the ultrasound imaging system's own image frames to track probe manipulations, making the system self-sufficient. The neural network processes existing image data that is already being captured for diagnostic purposes, eliminating the need for separate tracking systems and reducing overall complexity.
Solution Approach 2:
The neural network serves multiple functions: it identifies anatomical landmarks for diagnostic purposes and simultaneously tracks probe position and orientation for performance monitoring. This multi-functionality reduces the need for separate systems and minimizes added complexity.
3Loss of information
If log files include both user inputs and probe manipulations, then comprehensive performance records are generated, but log file size and processing load increase
Solution Approach 1:
The patent extracts only the essential probe manipulation data (position and orientation) from the comprehensive image analysis, separating this tracking information from the full diagnostic image processing pipeline. This extraction approach captures necessary performance metrics while avoiding redundant data storage.
Data Source
AI summary
Ultrasound image devices, systems and methods are provided. A system is configured to determine and track manipulation of an ultrasound probe during an ultrasound exam. Tracked probe manipulations may be synced with recorded user inputs received by the system during the exam. Probe manipulations paired with user inputs are stored as comprehensive log files documenting the exam, which are then converted into narrative text summaries for user review. A neural network is implemented to determine probe positions and orientations utilized during the exam based on landmarks identified in acquired image frames. User inputs received by the system are filtered such that only inputs indicative of exam performance are paired with the determined manipulations of the probe. External sensors are not necessary to track probe motion, but may be implemented to augment or replace certain system features.