東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

A flexible and efficient image retri...

Vu, Khanh.

Linked to FindBook

Google Book

Amazon

博客來

A flexible and efficient image retrieval system.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	A flexible and efficient image retrieval system./
Author:	Vu, Khanh.
Description:	131 p.
Notes:	Major Professor: Kien A. Hua.
Contained By:	Dissertation Abstracts International63-02B
Subject:	Computer Science -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3042983
ISBN:	0493568646

A flexible and efficient image retrieval system.
Vu, Khanh.

A flexible and efficient image retrieval system. - 131 p.

Major Professor: Kien A. Hua.

Thesis (Ph.D.)--University of Central Florida, 2002.

A prototype has been developed to demonstrate the feasibility and efficiency of the system. I conclude the dissertation with future research directions

ISBN: 0493568646Subjects--Topical Terms:

890869
Computer Science

A flexible and efficient image retrieval system.
LDR:03089nam 2200277 a 45 001 936607
005 20110510
008 110510s2002 eng d
020 $a 0493568646
035 $a (UnM)AAI3042983
035 $a AAI3042983
040 $a UnM $c UnM
100 1 $a Vu, Khanh. $3 1260339
245 1 0 $a A flexible and efficient image retrieval system.
300 $a 131 p.
500 $a Major Professor: Kien A. Hua.
500 $a Source: Dissertation Abstracts International, Volume: 63-02, Section: B, page: 0885.
502 $a Thesis (Ph.D.)--University of Central Florida, 2002.
520 $a A prototype has been developed to demonstrate the feasibility and efficiency of the system. I conclude the dissertation with future research directions
520 $a In this dissertation, I present a flexible image retrieval system that allows users to define queries of arbitrary shapes in a query-by-example environment. To realize this system, I have addressed a number of major issues, described as follows. (1) Query-by-example (QBE) is the most popular query model for content-based image retrieval (CBIR). A typical query contains not only objects of interest but also irrelevant image areas. The latter, referred to as noise, has limited the effectiveness of existing CBIR systems. I define noise-free queries (NFQs), which are composed of only relevant regions <italic> identified by the user at the query time</italic>. The challenge is how to precompute the feature vectors if we do not know the matching areas at database build time. I present a similarity model based on a sampling-based matching framework. The model handles NFQs effectively, and is robust with respect to scaling, translation, and semantic constraints of the matching objects. (2) To support large image datasets, I introduce a novel indexing technique for this new environment. The technique represents a unified solution to the following problems. (a) Since we cannot assume any shape for user-defined queries, the proposed indexing structure must be able to handle arbitrary-shaped queries. (b) It must be robust to scaling and translation. (c) Image similarity is typically determined by a large number of features. Current indexing approaches fail for high dimensional searching, a phenomenon known as the <italic>curse of dimensionality</italic>. An effective dimensionality reduction method has to be devised. (d) Many of the existing indexing techniques are not scalable. That is, the search time increases faster than the linear function of the data size. The proposed technique must be efficient for large data sets. (3) To facilitate the proposed indexing procedure, the core areas of NFQs have to be identified. To automate this task, it is necessary to propose an efficient algorithm that is able to detect the optimal core areas
590 $a School code: 0705
650 $a Computer Science $3 890869
690 $a 098
710 2 $a University of Central Florida $3 1260320
773 0 $t Dissertation Abstracts International $g 63-02B
790 $a 070
790 1 $a Hua, Kien A., $e adviso
791 $a Ph.D
792 $a 200
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3042983