Voice-Driven Structured Medical Report Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Structured medical reporting is considered less efficient than free text reporting, particularly due to the lack of seamless integration with voice recognition technology.
Innovation Solution
A method and device for generating part of a structured medical report by obtaining medical information, extracting predefined keywords, refining findings through iterative keyword extraction and modification, and using voice recognition to recognize keywords for efficient entry and modification of structured findings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If structured reporting is used with predefined findings and properties, then standardization and quality are improved, but efficiency and ease of operation deteriorate due to lack of seamless voice recognition integration
Solution Approach 1:
The patent introduces an intermediary system that translates free-text voice recognition output into structured medical report findings. This mediator converts the natural language output of voice recognition systems into the standardized format required by structured reporting, thereby maintaining both the efficiency benefits of voice recognition and the quality advantages of standardization.
Solution Approach 2:
The patent segments the structured report generation process into distinct components: initial finding extraction from voice text, refinement through iterative keyword extraction, and final structure assembly. This segmentation allows voice recognition to handle the initial content generation efficiently while systematic processing ensures proper structuring and standardization.
2Productivity
If free text reporting with voice recognition is used, then efficiency and ease of operation are improved, but standardization and quality deteriorate
Solution Approach 1:
The patent implements a dynamic refinement process where the structured finding is iteratively improved through multiple passes of keyword extraction and specification. The system starts with initial keywords from voice recognition and dynamically refines them through predefined rules and models, adapting the level of detail and structure based on the specific medical context and requirements.
Solution Approach 2:
The patent incorporates feedback mechanisms where the structured finding is continuously refined based on predefined rules, models, and tables that evaluate and improve the initial voice recognition output. This feedback loop ensures that the final structured report meets quality and standardization requirements while preserving the efficiency benefits of voice recognition.
3Manufacturing precision
If iterative refinement of findings is performed multiple times, then manufacturing precision and quality are improved, but loss of time increases
Solution Approach 1:
The patent performs preliminary actions by pre-defining rules, models, and tables that guide the keyword extraction and refinement process. These pre-established frameworks enable the system to quickly evaluate and refine findings without requiring time-consuming manual analysis at each iteration, thereby maintaining high accuracy while reducing overall processing time.
Solution Approach 2:
The patent changes parameters systematically during refinement, using predefined rules to adjust keyword specificity, finding structure, and detail level. By automatically adjusting these parameters based on the refinement stage and medical context, the system achieves high finding accuracy without requiring excessive refinement iterations, thus minimizing time loss.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The present invention relates to method and a device for generating part of a structured medical report. The method comprises a stepwise creation and refinement of findings related to an examined body part of a subject based on keywords extracted from medical information, such as words spoken by a user.