Evidential Reasoning in Expert Systems for Image Analysis.
Abstract
This report documents efforts to understand approaches to evidential reasoning of use in application of expert-system or knowledge-based-system techniques to image analysis (IA). There is growing evidence that these techniques offer significant improvements in image analysis, particularly in the coordinated application of specialized algorithms. This effort has four principal goals: (1) to clarify the basic issues in evidential reasoning (ER), (2) to provide a common framework for analysis, (3) to structure the ER process for major expert-system tasks in image analysis, and (4) to identify promising directions for further research. This research was carried out in three major segments. The first segment structured the evidential reasoning problem in a formal paradigm robust enough to be of practical use in design and construction of expert systems. It then formulated six important theoretical approaches in a parallel fashion in order to identify key assumptions, similarities, and differences. The second segment applied each of the ER approaches to three important tasks for expert systems in the domain of image analysis. This segment concluded with an assessment of the strengths and weaknesses of each approach. The third segment addressed promising directions for further research. It reviewed current results and identified important questions bearing on successful application of expert-system technology to image analysis.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1985
- Accession Number
- ADA153379
Entities
People
- T. R. Thompson
Organizations
- System Planning Corporation