Caption-Based Natural-Language Interface Handling Descriptive Captions for aMultimedia Database System

摘要

This research examined the grammar structure of descriptive English captions onmultimedia data. The research was composed of three phases. The first phase was to investigate the grammar structure of example descriptive captions from a variety of subject domains. The second phase was to develop a set of domain independent binary grammar rules to be used in the Caption-Based Interface (CBI) which is a natural language interface for the Multimedia Database System. The third phase of the research was to implement and test the grammar rules in the CBI. The program was implemented in C-prolog on a Sun SPARC workstation. The testing phase also includes timing and memory comparisons between C-Prolog an interpretive programming language and a compiled version of the code using Quintus Prolog. This thesis was able to show that the grammar rules that were developed could correctly identify their intended structures. Another accomplishment of this thesis was to demonstrate that the CBI could parse 25 out of 30 example captions and that it could correctly handle some semantic interpretations of the parsed captions.