City University of Hong Kong

CityU Institutional Repository >
3_CityU Electronic Theses and Dissertations >
ETD - Dept. of Computer Science  >
CS - Master of Philosophy  >

Please use this identifier to cite or link to this item:

Title: Content-based retrieval of 3D models using computational intelligence techniques
Other Titles: Li yong zhi neng ji suan li lun jin xing san wei mo xing de ji yu nei rong jian suo
Authors: Yeung, Pui Fong (楊佩芳)
Department: Dept. of Computer Science
Degree: Master of Philosophy
Issue Date: 2006
Publisher: City University of Hong Kong
Subjects: Computational intelligence
Image processing -- Digital techniques
Three-dimensional imaging
Notes: CityU Call Number: TA1637.Y48 2006
Includes bibliographical references (leaves 154-166)
Thesis (M.Phil.)--City University of Hong Kong, 2006
xiii, 166 leaves : ill. (some col.) ; 30 cm.
Type: Thesis
Abstract: Due to the rapid development in multimedia in the past few decades, many techniques and advanced approaches have been developed, but these techniques are mainly for the processing of 2D media entities such as images and videos. On the other hand, 3D media entities such as computer graphics models are becoming more and more important in multimedia, as can be witnessed by the wide adoption of 3D animation techniques in movies and computer games. In addition, 3D computer graphics models are also widely adopted in areas such as CAD/CAM, medical imaging and forensic applications. Due to the very different nature of 3D computer graphics models as compared with 2D media entities, new processing techniques are required for effective characterization of their properties. In view of this requirement, my research mainly focuses on the development of new and effective techniques for content-based retrieval of computer graphics models. Content-based 3D model retrieval is an emerging technique which is important for applications such as the rapid assembly of a virtual scene, in which a complete scene can be constructed from a partial setup by using a set of “seed objects” as queries to retrieve other relevant objects of the scene. While previous approaches focus on similarity retrieval, in which a 3D model is used as query to retrieve similar models in the database, our proposed approaches allow the specification of associative relationships which are more general than similarity between the query and database models. As a special example of this associative retrieval framework, we propose a content-based retrieval approach for 3D human head models based on a single 2D face view query. The motivation of adopting this particular approach and example is two-fold: 1. 3D human head models represent an important category of computer graphics models which are useful for constructing virtual characters and animated agents in multimedia applications. 2. Queries in the form of 2D face views are more readily available compared with 3D head models due to the wide availability of webcams. In addition, we also address the problem of heterogeneous associative retrieval of 3D models, in which the query model and the database models are from different categories but are closely associated with each other through a pre-specified relationship. This is unlike the case of conventional content-based 3D model retrieval applications, where the main objective is to retrieve a set of models which are similar in appearance to the query object. Due to the increasing need to perform rapid prototyping of virtual characters and rapid assembly of virtual scenes in entertainment media production, game design and virtual reality, simple and effective approaches for assembling and associating 3D models become more and more important. The proposed technique can serve as a unified approach for adaptively associating different models or model components, through which a complete model or virtual world could be constructed through an initial partial specification, instead of requiring a complete re-design each time.
Online Catalog Link:
Appears in Collections:CS - Master of Philosophy

Files in This Item:

File Description SizeFormat
fulltext.html159 BHTMLView/Open
abstract.html159 BHTMLView/Open

Items in CityU IR are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0!
DSpace Software © 2013 CityU Library - Send feedback to Library Systems
Privacy Policy · Copyright · Disclaimer