Convolutional neural network for content-based image retrieval system

Pao Yue-kong Library Electronic Theses Database

Convolutional neural network for content-based image retrieval system

 

Author: Zhao, Ruohan
Title: Convolutional neural network for content-based image retrieval system
Degree: M.Phil.
Year: 2016
Subject: Content-based image retrieval.
Image processing -- Digital techniques.
Neural networks (Computer science)
Hong Kong Polytechnic University -- Dissertations
Department: Dept. of Computing
Pages: 62 unnumbered pages : color illustrations
Language: English
InnoPac Record: http://library.polyu.edu.hk/record=b2890595
URI: http://theses.lib.polyu.edu.hk/handle/200/8468
Abstract: Content-based image retrieval (CBIR) is a combination technology, aiming to solve computer vision problems. The target is that digital images stored in database is requested by users in some way that based on content query. Comparing to other search analyzes, a significant differentiation is to exclude keywords and tags, only to focus on context itself. Because requesting people to type is time-consuming and can't obtain relative desirable picture, CBIR can promote efficiency and avoid subject. In this dissertation, I developed a website and IOS application together to finish CBIR. The image database contain all kinds of category. For my present demand, I collect featured representations along The Silt Road. A main technology is based on Convolutional Neural Network (CNN). CNN is a type of network where neurons tiled together in order to respond input (impulse). After training, extracting, comparing, match process finishes. There are overall around 2000 pictures and 7 category in database, users may choose to experience relative function through website. I also write Apple application where users may use mobile phone to retrieval image. ConvNet is based on Matlab code. System can learn database pictures and determine similar pattern. When users ask for a request, the server itself will analyze and process image data, then return demanded results. The performance is quite good to achieve image retrieval. Since it is different from traditional compare between two images, deep learning endows enough intelligence to machine for learning part. Thus, the outcome is supposed to be strongly objective. Response time is a matter of problem that should be refined in later research.

Files in this item

Files Size Format
b28905957.pdf 1.297Mb PDF
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.

     

Quick Search

Browse

More Information