Medical Image Report Generation With Automated Finding Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies struggle to easily support the generation of interpretation reports from medical images, particularly in detecting and specifying findings for medical imaging.
Innovation Solution
A document creation support apparatus and method that extracts regions with preset physical features from medical images, generates comments on findings using disease names associated with these features, and supports the creation of interpretation reports through user interaction and machine learning models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple processes are executed to detect abnormal shadows and specify findings for interpretation reports, then the detection accuracy is improved, but the operation complexity increases
Solution Approach 1:
The patent combines multiple separate processes (abnormal shadow detection, finding detection, and finding specification) into a single integrated interpretation report generation process. The processor automatically performs all these steps sequentially, merging what were previously separate operations into one unified workflow that reduces manual intervention while maintaining detection accuracy.
Solution Approach 2:
The system enables automatic generation of interpretation reports by having the processor self-perform the entire workflow from detecting abnormal shadows to specifying findings and generating the final report. This self-service mechanism eliminates the need for manual specification of findings by radiologists, allowing the system to autonomously complete the interpretation process while preserving diagnostic accuracy.
2Productivity
If automatic finding specification is implemented, then the productivity is improved, but the reliability decreases
Solution Approach 1:
The system incorporates feedback mechanisms where the processor automatically specifies findings based on detected abnormal shadows, generates interpretation reports, and allows for verification and correction. The feedback loop ensures that automatic finding specification maintains reliability by enabling review and adjustment of generated content, thus balancing productivity improvement with interpretation accuracy.
3Reliability
If manual specification of findings is required, then the reliability is maintained, but the loss of time increases
Solution Approach 1:
The system performs preliminary detection of abnormal shadows and automatic specification of findings before final report generation. By pre-processing the image data and automatically identifying potential findings, the system prepares the groundwork for report creation, significantly reducing the time radiologists need to spend on manual finding specification while maintaining interpretation reliability through subsequent review.
Data Source
AI summary
A document creation support apparatus comprising at least one processor, wherein the processor is configured to: extracts regions having one or more preset physical features from a medical image, and generates a comment on findings using a disease name associated with a physical feature of at least one of the extracted regions.