Medical Image Lesion Display for False Positive Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional medical image analysis systems using computer processing, such as CAD and AI, often produce false positive lesion detection regions, leading to annoyance for skilled doctors and potential oversight by unskilled doctors, as these false positives are repeatedly detected and displayed.
Innovation Solution
A system that distinguishes between false positive and non-false positive lesion detection regions by displaying them in different manners, using a hardware processor to acquire and display information on whether a region is false positive based on user input and previous analysis results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If computer processing is used to detect lesion regions in medical images, then detection speed and productivity are improved, but false positive regions are generated causing annoyance and potential oversight
Solution Approach 1:
The system performs preliminary action by storing information about false positive regions identified in previous image analyses. When a new image is analyzed, the system retrieves and compares this stored information to identify and flag regions that are likely to be false positives, preventing their misinterpretation before final diagnosis is made.
Solution Approach 2:
The system implements feedback by using the results of previous analyses (whether regions were determined to be false positives) to improve subsequent analyses. The stored information from past interpretations feeds back into the detection process, allowing the system to learn from previous errors and reduce false positives in future detections.
2Ease of operation
If all lesion detection regions are displayed uniformly, then the display is simple and easy to operate, but users cannot distinguish false positive regions from true lesions
Solution Approach 1:
The system applies local quality by displaying different visual characteristics for different types of lesion regions. False positive regions are displayed with distinct visual attributes (such as different colors, patterns, or annotations) compared to regions without false positive history, allowing users to quickly distinguish between them while maintaining overall display simplicity.
Data Source
AI summary
A non-transitory computer-readable recording medium storing instructions causing a computer to execute: acquiring information on whether each of one or more first lesion detection regions is false positive, the one or more first lesion detection regions being input by a user with respect to a first analysis result and obtained by computer processing on a first medical image of a patient; acquiring a second analysis result including one or more second lesion detection regions that are obtained by computer processing on a second medical image of the patient; and displaying a false correspondence region in a different manner from a manner of displaying a non-false correspondence region. The false correspondence region corresponds to a first lesion detection region indicated as being false positive, and the non-false correspondence region corresponds to a first lesion detection region indicated as being not false positive, in the acquired information.