Medical Image Reporting With Knowledge Graph Validation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical data processing systems face inefficiencies in generating image reports and correcting errors in medical text, and there is a need for improved methods to process and store unstructured medical data efficiently and accurately.
Innovation Solution
A method and system that utilize a knowledge graph to automatically generate image reports from medical images, correct errors in medical text based on error types and locations, and convert unstructured medical data into structured data for efficient storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual methods are used for generating image reports and correcting medical text, then accuracy can be maintained through expert review, but processing efficiency and productivity are significantly reduced
Solution Approach 1:
The system performs preliminary actions by pre-processing medical images to extract feature images and pre-processing medical text to identify entities and relationships before report generation. This prepares data in advance, enabling faster and more accurate automated processing without sacrificing quality
Solution Approach 2:
The patent introduces an intermediary automated processing system that acts as a mediator between raw medical data and final reports. This intermediary system uses AI models and knowledge graphs to bridge the gap between unstructured medical data and structured reports, significantly improving processing efficiency while maintaining accuracy through multiple verification layers
2Productivity
If automated systems are used for image report generation, then processing speed increases, but accuracy and reliability may be compromised without proper validation mechanisms
Solution Approach 1:
The system implements feedback mechanisms where processing results are continuously validated and refined. The automated processing system receives feedback from validation rules, knowledge graph consistency checks, and quality metrics, adjusting its output to maintain high accuracy while preserving processing speed benefits
Solution Approach 2:
Validation rules and quality checks are established in advance before automated processing begins. This preliminary setup ensures that accuracy requirements are built into the system architecture, allowing fast automated processing to proceed within predefined accuracy boundaries
3Ease of manufacture
If unstructured medical data is stored directly, then storage simplicity is maintained, but data retrieval efficiency and processing capability are reduced
Solution Approach 1:
The patent segments unstructured medical data into structured components during processing. Medical images are divided into feature images with extracted characteristics, and medical text is segmented into entities, relationships, and attributes. This segmentation enables efficient storage and rapid retrieval while maintaining ease of processing
Solution Approach 2:
The system changes the parameters of stored data from unstructured formats to structured formats with defined parameters and attributes. This transformation maintains storage simplicity through standardized schemas while dramatically improving retrieval efficiency and processing capability through organized data access
Data Source
AI summary
A method may include: obtaining a target image; obtaining a first feature image from the target image, the first feature image including a region of interest (ROI) in the target image; obtaining a second feature image that matches the first feature image; obtaining a knowledge graph, the knowledge graph including a relationship between the second feature image and image description information and diagnostic result information corresponding to the second feature image; obtaining the image description information and the diagnostic result information corresponding to the second feature image based on the knowledge graph and the second feature image; and generating an image report of the target image based on the target image, the image description information, and the diagnostic result information.