Due to the growing volume of digital image data in recent years, the demand and necessity for a system that can manage large image sets are rising. It is vital to create sophisticated retrieval functions in order to enhance the user’s convenience. A lot of attention is now being placed on feature-based similarity retrieval which employs features extracted automatically from images. We have been carrying out research on the image retrieval system called ExSight which makes use of feature-based similarity retrieval. In this paper, we propose a similarity measure for evaluating image shape similarity, and an access method for the proposed similarity measure. We also discuss our experimental observations.
Content
Author and article information
Conference
Publication date:
February
1998
Publication date
(Print):
February
1998
Pages: 1-13
Affiliations
[0001]NTT Information and Communication Systems Laboratories
Yokosuka, Kanagawa, Japan