SBIR-STTR Award

The objective of this research is to significantly reduce the human intervention required to link text and graphics. Many current research efforts have focused on linking text to graphics objects by: (1) developing textual graphics descriptors which can be used to identify graphics objects and (2) developing methods for linking, interpreting, and integrating text to graphics. This effort proposes to automate the identification of graphics objects and the creation of graphics descriptors which has traditionally been done manually. The research will complement and enhance the effectiveness of many existing research efforts by providing a means for automating many of the manual tasks associtated with integrating text and graphics. The specific objectives of this research are to automate the learning of fundamental graphics objects, to construct a knowledge base of these learned graphics objects, and to develop techniques for recognizing these objects within a complex graphic.

Keywords:
GRAPHICS OBJECTS AUTOMATED LEARNING RECOGNITION TECHNIQUES KNOWLEDGE BASE

Integration of graphics and text, as demanded in the authoring of maintenance procedures, requires the author have knowledge about the content of graphics files so the appropriate graphic can be located and matched with the maintenance text being written. Current attempts to assist authors in this difficult and time-consuming task include naming graphics files in such a way as to reflect their content. An automated search engine capable of searching files for specific graphics objects is what is needed. This Phase II proposal extends the results of van Houten Technologies Inc.'s successful Phase I work, to develop a graphics search engine, to produce the Intelligent Graphics Management Environment (IGME). The IGME functions similar to a human by "learning" graphics objects and then applying this knowledge to recognize previously unseen graphics. Artificial intelligence and machine learning technologeis are used to provide a powerful graphics search engine along with facilities to add, update, and delete graphics objects from the environment without requiring conversion of graphics file formats.

Keywords:
INTELLIGENT GRAPHICS SEARCH ENGINE GRAPHICS KNOWLEDGEBASES GRAPHICS AND TEXT INTEGRATION AUTOMATED