Author: Tam, Kam-shing
Title: Similar shape retrieval in image database system
Degree: M.Sc.
Year: 2001
Subject: Image processing
Hong Kong Polytechnic University -- Dissertations
Department: Multi-disciplinary Studies
Department of Electronic and Information Engineering
Pages: v, 51 leaves : ill. ; 30 cm
Language: English
Abstract: The ability to extract invariant features from an image is important in the field of pattern recognition and image processing. By using the invariant features, any object in the image can then be identified independently of its position, size and orientation. In this dissertation, different shape representation methods will be investigated and an appropriate one will be selected and evaluated. Furthermore, efficient methods, which combine the different shape features into a unified high-dimensional invariant feature set for discriminatory object retrieval, will be described. Experiments have been conducted based on a database consisting of 50 images taken from multicolored man-made or natural objects in real world scenes. A set of shape descriptors is firstly developed to measure the roughness of contours. A color image is then converted into a binary image. The respective contours in the image are then obtained and their corresponding shape descriptors are generated. These shape descriptors, including shape factor, invariant moments, and Fourier descriptors, are computed and evaluated for shape retrieval in this dissertation. Finally, their applicability for classification is studied by using the combination of these shape features. Furthermore, an image retrieval system will be constructed in order to access or retrieve images from a database based on a query image. At the moment, we will restrict ourselves to the case where an object is rigid and can be described by its boundary. Also, we will deal only with non-overlapping boundaries that are completely known (ideal input). We will assume, furthermore, that the classes can be described by a few fixed prototypes that may be rotated, translated and scaled. The objects can possess some random noise on their boundaries, such a situation may arise, for example, in the recognition of machine parts in automated assembly. The best classification result of 50 images from 5 classes using the combined shape features with the nearest-neighbor classification was 96%.
Rights: All rights reserved
Access: restricted access

Files in This Item:
File Description SizeFormat 
b15995732.pdfFor All Users (off-campus access for PolyU Staff & Students only)2.13 MBAdobe PDFView/Open

Copyright Undertaking

As a bona fide Library user, I declare that:

  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.

By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

Show full item record

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