Tissue Image Abnormality Detection With Over- and Underdetection Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing drug discovery methods face challenges in accurately detecting morphological abnormalities in tissue specimens due to over- or under-detection, which are not adequately addressed by current technologies.
Innovation Solution
A drug discovery support device and method that utilize a processor to detect, analyze, and provide cause identification reference information for over- or under-detection of morphological abnormalities in tissue specimens, using machine learning models and determination thresholds to enhance detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automatic image analysis is used to detect morphological abnormalities, then detection efficiency is improved, but detection accuracy deteriorates due to over- or under-detection
Solution Approach 1:
The system generates determination reference information based on detection results and feeds it back to identify causes of over- or under-detection. This feedback loop enables automatic adjustment and optimization of detection accuracy while maintaining high detection efficiency through continuous improvement without requiring manual expert intervention for each case
Solution Approach 2:
The system automatically identifies causes of detection errors and generates determination reference information without requiring external expert intervention. The automatic image analysis system serves itself by autonomously detecting morphological abnormalities, analyzing detection results, and generating reference information for improving its own detection accuracy
2Measurement precision
If manual observation by experts is used to detect morphological abnormalities, then detection accuracy is improved, but productivity deteriorates due to time-consuming analysis
Solution Approach 1:
The system introduces determination reference information as an intermediary between automatic image analysis and final detection results. This intermediary element enables automatic analysis to achieve expert-level accuracy by providing reference standards that guide and validate the automated detection process, eliminating the need for manual expert observation while maintaining high detection accuracy
3Measurement precision
If detection threshold is lowered to reduce under-detection, then detection sensitivity is improved, but false positive rate increases leading to over-detection
Solution Approach 1:
The system dynamically adjusts detection parameters based on determination reference information generated from actual detection results. By analyzing patterns in detection data and generating adaptive reference standards, the system optimizes detection sensitivity while maintaining appropriate false positive rates, avoiding the need for fixed threshold settings that cannot adapt to different tissue types and detection scenarios
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A drug discovery support device includes a processor. The processor is configured to obtain a specimen image of a tissue specimen of an organ of a subject subjected to an evaluation test of a candidate substance; detect, from among portions of the specimen image, one or more estimated morphological abnormality portions where a morphological abnormality is estimated to have occurred; derive determination reference information from a detection result of each of the one or more estimated morphological abnormality portions; determine, based on the determination reference information, whether the one or more estimated morphological abnormality portions have been overdetected and whether the one or more estimated morphological abnormality portions have been underdetected; in response to a determination being made that the one or more estimated morphological abnormality portions have been overdetected, present to a user first cause identification reference information contributing to identifying a cause of the one or more estimated morphological abnormality portions having been overdetected; and in response to a determination being made that the one or more estimated morphological abnormality portions have been underdetected, present to the user second cause identification reference information contributing to identifying a cause of the one or more estimated morphological abnormality portions having been underdetected.