Related Experiment Videos
Knowledge-based medical image analysis and representation for integrating content definition with the radiological
C A Kulikowski1, L Gong, R S Mezrich
1Department of Computer Science, Rutgers University, New Brunswick, NJ, USA.
Methods of Information in Medicine
|March 1, 1995
Summary
Integrating medical images and clinical reports requires compatible knowledge representation. A new knowledge-based method for medical image analysis facilitates automated comparison of image segmentation with expert interpretations.
Area of Science:
- Medical image analysis
- Artificial intelligence in healthcare
- Knowledge representation
Background:
- Advancements in computing enable integration of multimedia medical data for clinical decision-making, education, and research.
- Evaluating automated image segmentation requires correlation with expert radiological interpretations found in clinical records.
Purpose of the Study:
- To develop a method for ensuring compatibility between medical image content and textual descriptions in clinical records.
- To enable automated comparison of image segmentation results with expert judgments.
Main Methods:
- A knowledge-based, object-centered, hierarchical planning method for automated medical image analysis.
- Representing specialist problem-solving steps at the knowledge level (goals, tasks, objects, concepts).
- Separating knowledge representation from implementation details for diverse image types and analysis methods.
Main Results:
- The developed system generates structured image descriptions essential for matching with expert interpretations.
- Facilitates automated extraction of image content descriptions from text.
- Enables automated matching between image segmentation and textual descriptions.
Conclusions:
- The proposed knowledge-based approach is crucial for structured medical image analysis.
- This system can be a key component in building integrated hybrid patient data information systems.
- Enables systematic evaluation of automated image segmentation against expert radiological assessments.