Ultrasonic Imaging Validity Assessment via Neural Network Intermediary
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In ultrasonic imaging, neural networks trained by machine learning struggle to provide intuitive understanding of their behavior, making it difficult for users to determine the validity of images generated, especially for unknown inputs.
Innovation Solution
An ultrasonic imaging device is equipped with an image generation unit, a trained neural network, and a validity information generation unit that compares reception signals and images to generate validity information, allowing users to assess the validity of estimated images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a neural network is used for image reconstruction and quality improvement, then image quality and processing accuracy are improved, but it becomes difficult for users to determine the validity of the generated images
Solution Approach 1:
The patent introduces an explanation generation unit that creates intermediate explanatory information about the neural network's processing behavior. This intermediary element mediates between the neural network's internal operations and the user's understanding, providing insights into how input images are transformed into output images without requiring users to understand complex neural network mechanics.
Solution Approach 2:
The system implements feedback by generating explanatory information that returns to the user about the neural network's decision-making process. This feedback loop allows users to verify whether the network's transformations are reasonable and expected, thereby enabling validity determination while maintaining high image quality.
2Measurement precision
If machine learning techniques are applied for image reconstruction, then processing accuracy is improved, but the behavior of the system becomes difficult to intuitively understand
Solution Approach 1:
The explanation generation unit serves as an intermediary that translates the neural network's internal processing behavior into human-understandable explanations. It generates information about which regions of the input image are most important and how they are transformed, making the system's behavior intuitive without sacrificing processing accuracy.
Solution Approach 2:
The system uses visual representations (analogous to color changes) to highlight important regions and transformations in the image processing. By visually emphasizing key areas and changes, the system makes its behavior more intuitive and easier to understand while maintaining high processing accuracy.
Data Source
AI summary
An object of the invention is to provide a user with information that serves as a material for determining whether an image generated by processing including a neural network is valid. A reception signal output by an ultrasonic probe that has received an ultrasonic wave from a subject is received, and an ultrasonic image is generated based on the reception signal. A trained neural network receives the reception signal or the ultrasonic image, and outputs an estimated reception signal or an estimated ultrasonic image. A validity information generation unit generates information indicating validity of the estimated reception signal or the estimated ultrasonic image by using one or more of the reception signal, the ultrasonic image, the estimated reception signal, the estimated ultrasonic image, and output of an intermediate layer of the neural network.