Date of Award:

12-2013

Document Type:

Dissertation

Degree Name:

Doctor of Philosophy (PhD)

Department:

Computer Science

Committee Chair(s)

Xiaojun Qi

Committee

Xiaojun Qi

Committee

Stephen Clyde

Committee

Kyumin Lee

Committee

Vicki Allan

Committee

Yangquan Chen

Abstract

Digital imaging was a great invention in the last century. Since digital cameras became popular in the public, a large amount of digital images emerged in the late of the twentieth century. How to manage the huge amount of images and find desired images among them became an urgent issue during the same period.

Techniques of retrieving a desired image are generally categorized into two basic classes. One relies on text-based key words to retrieve desired images in the image
database. The other one relies on image-based queries to retrieve desired images in the image database. The second technique is usually named the content-based image retrieval technique. Major techniques involved in the content-based image retrieval technique include the image feature extraction, the feature matching algorithm, and the similarity calculation. Each technique plays an important role in the content-based image retrieval, and they have their own challenge issues as well. For instance, how to find an efficient and accurate feature matching algorithm is still a hot topic in the content-based image retrieval.

This dissertation addresses certain challenge issues that exist in the content-based image retrieval technique and proposes two different retrieval systems that can be applied in the small-scale and the large-scale image databases.

Checksum

446c967025b47ed820632d1961401172

Share

COinS