From nobody@cse.Buffalo.EDU Wed May  5 12:02 EDT 1999
From: Nobody <nobody@cse.Buffalo.EDU>
Date: Wed, 5 May 1999 12:02:23 -0400 (EDT)
To: techreps@cse.Buffalo.EDU
Subject: techrep: POST request
Content-Type: text
Content-Length: 1238

ContactPerson: azhang@cse.buffalo.edu
Remote host: merope.cse.buffalo.edu
Remote ident: azhang
### Begin Citation ### Do not delete this line ###
%R 99-04
%U /multimedia/azhang/pub/acm99.ps
%A Yu, D. 
%A Zhang, A.
%T ACQ: An Automatic Clustering and Querying Approach for Large Image Databases
%D May 05, 1999
%I Department of Computer Science and Engineering, SUNY Buffalo
%K data mining
%X Large image collections such as web-based image databases are being 
built in various locations. Because of the diversity of such 
image data collections, clustering images becomes 
an important and non-trivial problem. Such clustering tries to find  
the densely populated regions in the feature space to be used for 
efficient image retrieval. 
In this paper, we present an automatic clustering and querying ACQ 
approach  
for large image databases. Our approach can efficiently detect clusters of 
arbitrary shape. It does not require the number of clusters to be 
known a priori and is insensitive to the noise (outliers) and the 
order of input data. Based on this clustering approach, efficient 
image querying is supported. 
Experiments demonstrate the effectiveness and efficiency 
of the approach.