AI Medical Image Analysis Worklist Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current AI-based medical image analysis technologies do not efficiently differentiate between normal and abnormal images, leading to an unchanged workload for radiologists as they still need to review all images, including normal ones.
Innovation Solution
An AI-based medical image analysis method and system that uses an AI model to detect indeterminate or abnormal regions in medical images, classify images as normal, indeterminate, or abnormal, and automatically generate reports for normal cases, thereby prioritizing abnormal cases for reading.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AI-based medical image analysis is implemented to detect abnormalities, then the accuracy and reliability of abnormal case identification is improved, but the overall workload for radiologists remains unchanged as normal images still require review
Solution Approach 1:
The patent segments the medical image review workload by classifying images into normal and abnormal categories using AI analysis results. The worklist is divided into separate sections showing only abnormal cases that require radiologist review, while normal cases are automatically filtered out. This segmentation allows radiologists to focus exclusively on abnormal cases, improving reading efficiency without compromising abnormality detection accuracy.
2Reliability
If radiologists review all images including normal ones, then comprehensive coverage is ensured, but time is wasted on images that do not require detailed analysis
Solution Approach 1:
The system performs preliminary AI-based analysis and classification of medical images before they are presented to radiologists. By pre-identifying and flagging abnormal cases, the system prepares the worklist in advance to show only cases requiring radiologist review. This preliminary action ensures that comprehensive coverage is maintained for abnormal cases while eliminating time waste on normal cases.
3Measurement precision
If AI model detects suspicious regions with high sensitivity, then abnormal cases are reliably identified, but false positives increase requiring additional verification
Solution Approach 1:
The patent introduces an intermediary verification mechanism where the AI system provides structured analysis results and confidence scores for detected suspicious regions. These intermediate results serve as a bridge between raw AI detection and final radiologist diagnosis, allowing radiologists to quickly assess the reliability of AI-detected abnormalities and prioritize verification efforts accordingly.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image analysis device includes a memory and a processor configured to execute instructions stored in the memory, wherein the processor is configured to detect an indeterminate region or an abnormal region from an input medical image using an artificial intelligence (AI) model trained to detect a suspicious region and a lesion region in medical images and determine the input medical image as a normal case when the indeterminate region or the abnormal region is not detected in the input medical image.