Failed-Image Decision Support With Explainable Imaging Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical image analysis systems fail to provide users with clear explanations for why an image is determined as failed, making it difficult for users to accept and understand the determination results.
Innovation Solution
A failed-image decision support system that performs multiple types of failed-image determination processes on medical images, generating both a determination result and the basis for the result, and outputs this information to assist users in understanding the decision-making process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple types of failed-image determination processes are performed on a medical image, then the comprehensiveness of the determination result is improved, but the user's understanding of the determination basis deteriorates
Solution Approach 1:
The patent segments the determination basis information by associating it with specific determination processes. Each determination process (e.g., positioning assessment, image quality evaluation) has its own basis information, allowing users to trace which process contributed which conclusion. This segmentation maintains comprehensiveness while improving understandability.
Solution Approach 2:
The patent introduces determination basis information as an intermediary between the multiple determination processes and the final determination result. This intermediary element translates complex multi-process outcomes into interpretable explanations, bridging the gap between comprehensive analysis and user understanding.
2Ease of operation
If only the determination result and image are displayed, then the simplicity of the interface is improved, but the user's acceptance of the determination result deteriorates
Solution Approach 1:
The patent prepares determination basis information in advance alongside the determination result, so that when the result is presented to the user, the explanatory basis is already ready and integrated. This preliminary preparation ensures that acceptance-enhancing information is available without requiring complex real-time generation or separate retrieval steps.
3Loss of information
If detailed determination basis information is provided, then the user's understanding is improved, but the complexity of the output system deteriorates
Solution Approach 1:
The patent merges the determination result and determination basis information into a unified output structure. Rather than treating them as separate components requiring separate processing and display systems, they are combined in a way that the basis information naturally supports and explains the result, reducing overall system complexity.
Data Source
AI summary
A failed-image decision support apparatus includes a hardware processor and an outputter. The hardware processor performs a failed-image determination process on a medical image to thereby generate (i) a determination result and (ii) a guidance image for proper imaging. The failed-image determination process is a process using an inference obtained by a learned model, and the outputter outputs the determination result and the guidance image before re-imaging is performed.