画像-言語間を統括的に扱う知識表現法による対象に依存しない柔軟な画像知識学習

Translated title of the contribution: The flexible knowledge constructing using the integrated knowledge description of image and language.

堀口直孝, 荒井 秀一, Shuichi ARAI

Research output: Contribution to journalMisc

Abstract

For realizing flexible image understanding that is independent of objects,it is effective that a system acquires the knowlede by learning.There are two ways to learn,larning by instruction and automatic leaning.For actualizing both of ways,the knowledge for image understanding must be accessible by both of a man and a computer.On the other hand,the svstem must be able to handle all of the information from quantitative information of image to quailtative one about concept,even if the level of the knowledge changes by learning.Therefore,on this research.,we propose a new method to describe the knowledge.It can describe concepts efficiently and information of image on computers organically. Besides we constructed the image understanding system,and we confirmed the effect of the method by experiments.
Translated title of the contributionThe flexible knowledge constructing using the integrated knowledge description of image and language.
Original languageJapanese
Pages (from-to)AI94 - 28
JournalIEICE technical report. Artificial intelligence and knowledge-based processing
StatePublished - 1994

Fingerprint

Dive into the research topics of 'The flexible knowledge constructing using the integrated knowledge description of image and language.'. Together they form a unique fingerprint.

Cite this