Design and Implementation of COIRS (a COncept-based Image Retrieval System)

Authors:
Hyungjeong Yang;Hoyoung Kim;Jaedong Yang
Affiliations:
-;-;-
Venue:
VISUAL '99 Proceedings of the Third International Conference on Visual Information and Information Systems
Year:
1999

Citing 4
Cited 0

Retrieval of similar pictures on pictorial databases

Pattern Recognition
Photobook: content-based manipulation of image databases

International Journal of Computer Vision
An Object-Based Information Retrieval Model: Toward the Structural Construction of Thesauri

ADL '98 Proceedings of the Advances in Digital Libraries Conference
A letter oriented minimal perfect hashing function

ACM SIGPLAN Notices

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we design and implement COIRS (Concept-based Image Retrieval System). It is a content-based image retrieval system to search for images as well as indexing them based on concepts. The concepts are detected by a thesaurus called triple thesaurus. The triple thesaurus consists of a series of rules defining the concepts. COIRS adopts an image descriptor called triple to specify the spatial relationships between objects in an image. An image is indexed by a set of triples - each of them is enrolled into an inverted file, pointing to the image. We also develop a query processor to retrieve relevant images by evaluating a user query. The query formulated in terms of triples is evaluated by matching its triples with those of the inverted file.