From text to images through meanings

  • Authors:
  • Jayakumar S. Kandhasamy;Ming Xie

  • Affiliations:
  • Nanyang Technological University, Singapore;Nanyang Technological University, Singapore

  • Venue:
  • Proceedings of the 2004 ACM SIGCHI International Conference on Advances in computer entertainment technology
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Natural language is an easy and effective medium to interact with computers by humans and we foresee that it is an effective way to visualize or realize one's imagination by computer. And it makes the complex task of creating 3D scenes or images faster and easier. As a result, creation of 3D scenes will be seen as just programming the computer with human language such as English. This paper presents a meaning centric framework for text to 3D image system to visualize the meaning of text and we propose an active memory, which is based on object oriented programming technique, to construct the mental world or active database. Thus each word in the active database acts as an agent compared to just a graphic based annotation, which can be found in existing systems. The active memory incrementally acquires and organizes the lexicon of natural language, visual objects and behaviors through real time interaction with the environment (including human masters). Based on the meaning centric framework, we have started the development of a software agent called KnowNet. KnowNet is an on-line learning, tutoring and visualization system for English language, which supports bi-directional learning activities such as human learning from computer and computer learning from human. It uses VRML browser to render 3D images. Most importantly, KnowNet acts as an entertainer for people who is proficient in English language and it act as an English language teacher for those who wants to learn English language.