Abstract
A computerized expert system to determine the skeletal maturity from a hand-wrist radiograph is being developed, and we report progress so far. The advantages of the system are speed and total objectivity.
The two basic tasks associated with this problem involve image processing and artificial intelligence. As regards image processing, we have developed a computationally cheap method to identify each bone and to store its coordinates. It is based on simple shape descriptions and structural locations and is immune to changes in orientation and to growth related changes. It is encoded as a sub-expert system.
The second task involves the analysis of the characteristics of each bone to determine its maturity stage. We have developed rules based on the stage descriptions of Tanner et al, 1983 (Assessment of Skeletal Maturity and Prediction of Adult Height, 2nd ed., Academic Press). Each description is reformulated into a set of questions, and production rules are used to encode these as knowledge sources. The system architecture is similar to the blackboard structure used in HEARSAY-2. The knowledge sources correspond to every node of a classification tree and are invoked either from the blackboard or from another sub-expert system.
Article PDF
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Kwabwe, A., King, R. & Tanner, J. Progress towards a computerized expert system for the assessment of skeletal maturity. Pediatr Res 18, 1218 (1984). https://doi.org/10.1203/00006450-198411000-00106
Issue Date:
DOI: https://doi.org/10.1203/00006450-198411000-00106