A New Semantic Network Program Based on Combination of Case Knowledge and General Knowledge

This paper proposes a new semantic network program that combines case knowledge with well-defined general knowledge. The proposed semantic network program has been applied to scene understanding problem. When the network receives a scene image, it outputs candidates for an ambiguous image object in the scene image based on activation spreading on the network. Java programming language has been employed to program the semantic network. It has been shown that object-oriented principle fits semantic networks and makes them a practical tool to realize human-like information processing. Simulation of image object identification has been carried out and has shown that the proposed semantic network program can be an effective component in modeling interaction between logical and pattern information processing.