| Abstrakt: |
As the number and size of image databases grows, accurate and efficient content-based image retrieval (CBIR) systems become increasingly important in business and in the everyday lives of people around the world. Accordingly, there has been a substantial amount of CBIR research, and much recent interest in using probabilistic methods for this purpose. However, there exist some shortfalls or limitations of these methodologies. Methods which boost retrieval performance by incorporating knowledge base inference engine have also been of interest. In this paper we describe a novel framework for performing content-based image retrieval using knowledge based inference engine. Given a user specified image query, the system first interprets the query using the knowledge based inference engine and then extracts a set of images, from a labelled corpus, corresponding to that query. The experimental results reveal that the performance of image retrieval can be surprisingly enhanced. [ABSTRACT FROM PUBLISHER] |
