Medical Image Registration via Multi-Scale Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical image diagnostic techniques face challenges in registering ultrasound images with CT or MR images due to the difficulty in selecting initial positions and the time-consuming nature of the registration process, which is dependent on accurate feature extraction and requires multiple user inputs, leading to inefficiency and potential inaccuracies.
Innovation Solution
A medical image diagnostic apparatus that includes extraction circuitry for identifying structures, scale setting circuitry for adjusting display scales, and image registration circuitry for performing pre-registration and formal registration on multiple scales, allowing for automatic registration from coarse to fine levels and user correction inputs to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual feature extraction and point input are performed by the user, then registration accuracy can be improved through precise input, but the time required for registration increases significantly
Solution Approach 1:
The system performs automatic feature extraction and initial registration before requiring user input. The extraction circuitry automatically identifies features and performs preliminary registration, so that when the user provides correction inputs, the system is already close to the final result, reducing the total time required.
Solution Approach 2:
The system performs self-service through automatic feature extraction and preliminary registration. The extraction circuitry and registration circuitry automatically process the images without requiring continuous user intervention, only providing corrections when necessary.
2Measurement precision
If multiple user inputs are required to confirm registration demand, then registration accuracy can be improved, but the operational complexity and time consumption increase
Solution Approach 1:
The system provides feedback by displaying the registration result after each automatic registration attempt. The user can review the result and provide correction inputs only when necessary, creating a feedback loop that maintains accuracy while minimizing operational complexity.
Solution Approach 2:
The system performs preliminary automatic registration before requiring user confirmation. This preliminary action establishes a baseline that is often sufficient, reducing the number of times the user needs to intervene.
3Productivity
If automatic feature extraction is used, then registration speed improves, but registration accuracy becomes dependent on extraction precision
Solution Approach 1:
The extraction circuitry performs self-service by automatically extracting features without requiring user input. The system extracts features autonomously and uses them for preliminary registration, only requiring user correction when the automatic extraction is insufficient.
Solution Approach 2:
The system performs preliminary automatic feature extraction and registration before requiring user input. This preliminary action establishes a foundation that speeds up the overall process while maintaining accuracy through subsequent user corrections when needed.
4Measurement precision
If the user manually selects initial positions for registration, then registration accuracy can be improved, but the difficulty of operation increases due to the three-dimensional nature of CT images
Solution Approach 1:
The system performs preliminary automatic feature extraction and initial registration before requiring user input. This preliminary action handles the difficult initial position selection automatically, making the system easier to operate while maintaining accuracy through user corrections when necessary.
Data Source
AI summary
A medical image diagnostic apparatus according to an embodiment includes processing circuitry. The medical image diagnostic apparatus performs image registration between medical image data. The processing circuitry extracts a structure of the subject included in the medical image data. The processing circuitry sets a display scale of the subject to a specified value. The processing circuitry performs image registration between the medical image data on the display scale of the specified value.