Ultrasound Image Analysis Switching Between Internal and External Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ultrasound diagnostic apparatuses may not have high-performance analytic models due to structural or operational constraints, necessitating a mechanism to temporarily or supplementarily use external analytic models when internal models are not functioning adequately.
Innovation Solution
An ultrasound diagnostic apparatus and system that includes an internal analytic model and an external analytic model, with an evaluator to determine when to switch between them based on evaluations of internal data and output data, using saliency maps and confidence degree histories to ensure optimal performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If an internal analytic model is provided inside the ultrasound diagnostic apparatus, then responsiveness and operational speed are improved, but analysis accuracy and performance may deteriorate due to structural constraints
Solution Approach 1:
The patent introduces an evaluator as an intermediary component that assesses the quality of ultrasound images and determines whether to use the internal or external analytic model. This mediator resolves the contradiction by selecting the appropriate model based on image quality, ensuring both responsiveness (when internal model suffices) and accuracy (when external model is needed).
Solution Approach 2:
The system dynamically switches between internal and external analytic models based on real-time evaluation of image quality. This dynamic adaptation allows the system to optimize between speed and accuracy depending on the specific imaging conditions, rather than being fixed to one model.
2Measurement precision
If a high-performance external analytic model is used, then analysis accuracy is improved, but device complexity and operational overhead increase
Solution Approach 1:
The system employs partial use of the external analytic model only when necessary, as determined by the evaluator. Rather than always using the high-performance external model, the system applies it selectively to cases where it is needed, reducing overall system complexity while maintaining accuracy when required.
Solution Approach 2:
The internal analytic model serves itself by handling the majority of analysis tasks independently. The external model acts as a supplementary resource that is only invoked when the internal model's assessment indicates insufficient image quality, allowing the system to maintain simplicity while having access to high-performance capabilities.
3Ease of operation
If the internal analytic model is always used, then operational simplicity is maintained, but reliability deteriorates when image quality is insufficient
Solution Approach 1:
The evaluator provides feedback on image quality metrics and uses this information to automatically determine whether to switch to the external analytic model. This feedback mechanism ensures reliability by detecting when the internal model's performance may be insufficient and triggering appropriate model switching without user intervention.
Solution Approach 2:
The system prepares for potential performance issues by having the external analytic model available as a pre-arranged backup. When the evaluator detects insufficient image quality, the external model is already positioned to take over, providing a cushion against reliability failures without requiring complex real-time decision-making.
4Measurement precision
If the external analytic model is frequently used, then analysis performance is improved, but loss of time due to model switching increases
Solution Approach 1:
The evaluator performs preliminary assessment of image quality before initiating the analysis process. By evaluating image quality in advance and making model selection decisions beforehand, the system minimizes the time lost during actual model switching, as the decision is already made when the analysis begins.
Data Source
AI summary
An evaluation unit evaluates performance of an analytic model (internal analytic model) based on internal data extracted from the analytic model and output data of the analytic model. A determination unit determines use of an external analytic model based on an evaluation result of the evaluation unit. In that case, an input image (tomographic image) is sent to an information processing apparatus via a communication unit. An analysis result of the analytic model (external analytic model) is sent to a calculation control unit via the communication unit.