Study on Formalized Relevance Knowledge for Diagnosing Image Data: Toward Effective Distribution of Technical Know-How among Engineers

摘要

This paper presents a new approach to develop a know-how distribution support system that aims at facilitating the effective sharing of a variety of technical know-how among engineers, especially focusing on the knowledge exploited to diagnosing case data. By using annotations other diagnosticians attached to diagnosed images as clues, the proposed computational support tries to explore cases or images possibly related to the image an engineer is currently analyzing, so as to provide the user with new opportunities where he/she can acquire some analytical skills on his/her own. The proposed approach consists of two networks: the one is constructed based upon the statistical (co-occurrence) relationships involved in the set of annotated data; and the other is extracted from the respective human users by means of the DEMATEL method.