Use this URL to cite or link to this record in EThOS:
Title: Pictorial representation of text : converting text to pictures
Author: Ludlow, Nelson David
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 1993
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
'A picture is worth a thousand words.' This saying suggests that an inter-relationship exists between text and pictures. This thesis is the result of an investigation to identify and exploit the inter-relationships between text and pictures. It describes a concept of pictorial representation of text, presents a Text-to-Pictures System which generates a pictorial representation of English sentences, and gives a more detailed look at how the system pictorially represents specific linguistic types of natural language expressions. Although very little previous work has been done in this area, the relevant work in text-pictures systems is summarized. Most of the past work concentrated on pictorially representing the nouns and some spatial prepositions. My work expands the pictorial representation to include temporal expressions, conjunction, relative clauses, quantification, and some verb features. This thesis also addresses the concepts of using pictorial representation for data fusion of large natural language texts, as well as the problems of ambiguity and vagueness. The working system to 'convert text to pictures' is demonstrated. The system structure, intermediate representation schemes, and the translation process are described. Several types of natural language expressions are examined and the corresponding pictorial representations are shown. Also shown is an application to pictorially represent all of the possible meanings of an ambiguous sentence to allow a non-linguist user to choose the intended meaning. Using a natural language text processing system that can convert a sentence of text into a logical form (LF) representation, I show what is required to convert the LF representation into a pictorial representation. The process involves identifying the objects contained in the LF and representing them by icons. These icons are placed in an imaginary space via a set of constraints. After all of the constraints are determined, the system attempts to solve the generated constraint satisfaction problem. If a solution is found, the icons are drawn with the appropriate coordinates on a graphics display.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available