Computer-Aided Detection System with User-Validated Feature Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computer-aided detection (CAD) systems face challenges in accurately identifying and diagnosing abnormalities in medical images, often requiring radiologist expertise and leading to unnecessary biopsies due to inadequate automated detection and diagnosis capabilities.
Innovation Solution
A CAD system that segments medical images to extract relevant features, allows user interaction for confirmation or modification of detected features, and computes a diagnosis using a combination of pre-defined criteria and AI rules, providing a dynamic and validated diagnosis report.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If automated detection and diagnosis capabilities are enhanced, then diagnostic accuracy and reliability improve, but system complexity and requirement for user interaction increase
Solution Approach 1:
The CAD system segments the diagnostic process into distinct modules: image acquisition, preprocessing, feature extraction, analysis, and diagnosis generation. Each module handles specific tasks independently, allowing the system to manage complexity through functional decomposition while maintaining high diagnostic accuracy through specialized processing at each stage.
Solution Approach 2:
The system introduces an intermediary review stage where detected features and preliminary diagnoses are presented to users for validation and modification before finalization. This intermediary layer bridges automated detection and final diagnosis, enabling the system to leverage AI capabilities while incorporating human expertise to resolve complexity and ensure reliability.
2Productivity
If automated detection capabilities are improved, then productivity increases, but measurement precision of abnormalities may deteriorate
Solution Approach 1:
The system implements feedback mechanisms where detection results are continuously refined based on user validation and modification. Detected features are presented to users who can confirm or correct them, and this feedback loop continuously improves measurement precision while maintaining high productivity through automated initial detection and structured review processes.
Solution Approach 2:
The system performs preliminary automated detection and feature extraction before user review, preparing candidate abnormalities and extracted features in advance. This preliminary action handles routine detection tasks automatically to maintain productivity, while user review focuses specifically on validating and refining measurements to ensure precision.
3Reliability
If user interaction and validation are incorporated, then diagnosis reliability improves, but time consumption increases
Solution Approach 1:
The system implements partial user interaction by presenting only the most uncertain or critical detected features and diagnosis candidates for user review, rather than requiring validation of all findings. This partial action approach maintains diagnosis reliability for critical cases while reducing time consumption by automating confidence-high detections without user intervention.
Data Source
AI summary
The invention provides a system and method for computer-aided detection (“CAD”). The invention relates to computer-aided automatic detection of abnormalities in and analysis of medical images. Medical images are analyzed, to extract and identify a set of features in the image relevant to a diagnosis. The system computes an initial diagnosis based on the set of identified features and a diagnosis model, which are provided to a user for review and modification. A computed diagnosis is dynamically re-computed upon user modification of the set of identified features. Upon a user selecting a diagnosis based on system recommendation, a diagnosis report is generated reflecting features present in the medical image as validated by the user and the user selected diagnosis.