Mixed Query Image Retrieval System Bingjing Cai☆, Chris Zheng※, Sen Yang*, Jeffery Z. J. Zheng﹠ ☆
School of Software, Yunnan University, Kunming, Yunnan, P.R. China, 650091 Email:
[email protected]
※
Conjugate Systems Pty Ltd, 45 Greenways Road, Glen Waverley, Victoria 3150, Australia Email:
[email protected]
*
Department of CS&T, School of Information, Yunnan University, Kunming, Yunnan, P.R.China, 650091 Email:
[email protected]
﹠
School of Software, Yunnan University, Kunming, Yunnan, P.R.China, 650091 Email:
[email protected]
Abstract - This paper discusses a proposed mixed query
by keyword or by content. The first method was introduced
image and text retrieval database, combining text-based search
with the development of text-based search technology; the
technology and image content-based search technology for more
second method introduced with the content-based search
accurate results. The target application will be aimed at large,
technology.
categorized image sets, such as large image collections of libraries
Keyword-based image retrieval systems are based on the
or patent offices. It can be shown that this mixed query image
traditional text-based retrieval technology. The concept is
search engine, can achieve better efficiency and higher quality
simple. The images are tagged with keywords and managed by
than traditional text-based queries and content-based systems.
the system. Users search for images via a textbox and images
This paper will discuss the designing principle of the mixed query
tagged with the keyword will be displayed. The search is easy
image search engine, outline the architecture and show results
but the tagging becomes laborious. Keywords have to be
from an initial prototype database.
manually tagged to each picture by the user. For most people, this system of organizing is much too time consuming. If
Index Terms - content-based image search technology,
pictures are not tagged, then they cannot be organized and
keyword-based image search technology, mixed query image
data is lost. Furthermore, text labels do not represent the
retrieval system
picture itself, only a very abstract description which differs I.
A.
INTRODUCTION
Two Phrases of Development of Image Retrieval System Technology In the sudden explosion into the information age, all types
of data are being produced at an enormous rate and are still increasing rapidly. These data includes large sets of image, sound, video and other multimedia data being generated by cheap digital capture and storage facilities. With increase in demand for processing and categorized this rich multimedia information, large amounts of data are being created without a way to classify or search through them. It is with this need in mind that image search engines have come to be one of the hottest and fastest growing areas of research and application development. Currently, there exists two ways of searching for an image –
between people and cultures. There is no standard for tagging available and the only dependable tagging comes from ‘social tagging’ – an example being Flickr [1]. However, tag still cannot represent the content within and the search results will be unsuitable in many cases. Content-base search uses queries in the form of image objects
rather
than
text
tags. The
principle
behind
content-based retrieval technology is to extract either meaning or measurement from within the picture itself. Image properties such as color, texture, shape and other qualities can be expressed as a given measurement, able to be processed by the computer. Once the picture is able to be quantitatively and qualitatively defined, it thus can be managed. As a consequence, the real attraction for content-based search is that it promises the possibility for automation of image
classification and search.
systems, whether it is keyword-based search technique or
B.
image content-based search technique, cannot always obtain
Research and development of image retrieval techniques With the emergence of large scale image collections,
satisfactory results and meet users’ requirement. Effective and
content-based image retrieval was proposed. Since then, many
efficient system architecture for the image retrieval system is
techniques in this research direction have been developed and
needed, combining with both text-based search and image
many image retrieval systems, both research and commercial,
content-based search.
have been built. In the early 90s, IBM developed its first
Therefore, in this paper, we discuss a method of improving
content-based image retrieval, QBIC, standing for Query By
image retrieval for large image databases. It combines
Image Content system [2]. It is the first commercial
keyword-based search and image content-based search
content-based Image Retrieval system and its system
technologies. We name it as Mixed Query Image Retrieval
framework and techniques have profound effects on later
system. We present the system architecture, and illustrate our
Image Retrieval system. Photobook [3] is a set of interactive
system prototype. The experiment demonstrated that, the
tools for browsing and searching images developed at MIT
degree of accuracy, effect and efficiency of this mixed query
Media Lab. Photobook consists of three sub-books, from
image retrieval system are greatly enhanced. II.
which shape, texture, and face features are extracted
THE DESIGN PRICIPLE OF MIXED QUERY IMAGE
respectively. Users can then query based on corresponding
RETRIEVAL SYSTEM
features in each of the three sub-books. There are other
Recent years, there has been a rapid increase in the size of
content-based image retrieval systems such as the ViualSEEK
digital image collections. Both military and civilian equipment
system [4] developed at Columbia University, Like.com
generates gigabytes of images every day. As for such large
website system [5] developed by the Riya team in the United
size image databases, although image content-based search is
States and so forth.
faster and more accurate than text-based search, the retrieval
Like.com is one of the best true visual search engines, the
results are not always ideal. It is very possible that, for
contents of photos are used to search and retrieve similar items.
example, when we use the image content-based search engine
Its launch focuses on handbags, jewelry, shoes, and watches,
to seek a picture of a red bus, the return picture may be of a
allowing users to search and purchase items from thousands of
red house. As it is known to us, the image content-based
leading and boutique brands. It has classified image databases,
search engine extracts visual features from images, such as
such image databases of handbags and jewelry separately, and
color, texture and shape, according to different feature
has evident effect of content-based image retrieval.
extraction algorithms. It does not have human perception and
Currently, however, the popular model for image retrieval
the ability to distinguish and identify true meanings of images.
of most systems has been based on text-based search
The color and shape, as well as the texture of the house may
technology so far, such as Google, AltaVista, and Yahoo and
be very similar to the red bus. But it is not the result we expect.
so forth. Since the low-level features of images (color, shape,
But if in a relatively small, categorized image database, image
texture, etc) do not represent the image semantic information,
content-based retrieval engine definitely works better.
which means they do not tell what the image is, the results of
The design principle of the mixed query image retrieval
the content-based image retrieval are not always satisfying.
system is that first we divide the original large image database
C.
into a number of relatively small image databases based on
Existing problem and the objective of this paper Currently, most people focus on the improvement and
categories and establish keyword-indexing for each category
optimization a certain kind of content-based image retrieval
of the now divided image databases. We then employ image
techniques. Improvement of image content-based search
content-based search engine within each of the classified
techniques is definitely of great significance to obtain accurate
image database.
retrieval results in large size image databases. However, only
A.
one kind of image search techniques applied to image retrieval
Combination of Keyword-based Search and Image Content-based Search
There are two layers of combination of text-based and image content-based search. Firstly, according to categories, divide the original large image database into small image database
and
establish
keyword-on-category
indexing
associated with small databases. Secondly, in every classified image databases, establish both image content-based indexing and keyword-based indexing. 1)
Classify image database. As for large image database, we can refer to image content
information or descriptive information of images to classify the image database based on category. A number of small image sub-databases are generated and each of them belongs to a certain category, for instance, shoes or handbags. Take the image database of the patent office for example; suppose that the number of patent pictures is about 5,000,000 and the number of the patent categories is more than 40. If we divide
Fig.2. An example of classified image databases structure
the image database according to the standard international
Then, according to these categories, we can employ
patent category, a number of image sub-databases are
traditional text-based search indexing technique to establish
generated, which are classified and contain about 100,000 to
text index for the cluster of image databases. Thus, we could
200,000 images each. In smaller image database, both the
utilize key words to search for a certain kind of image
image content-based and text-bases search engines can obtain
collections.
more accurate retrieval results and work faster. And efficiency
2)
of the whole system can be enhanced and higher quality of search results can also be obtained.
Establishing Content-based Indexing After classifying the original large size image databases into
relative small image databases on different categories, we
A distributed cluster of sub-databases are generated after
could utilize image content-based search. Since we have
classifying the original large image databases. We can see it as
divided large size image database into small, categorized
shown in figure 1.
sub-databases, content-based search engine is more accurate
After classifying, the structure of image databases is similar
and faster. Feature (content) extraction is the basis of
to the category structure of libraries’ databases or patent office
content-based image retrieval. In a broad sense, features may
IPC categories structure [5]. Figure 2 shows an example of
include both text-based features (keywords, annotations, etc.)
such kind of categorized structure.
and visual features (color, texture, shape, faces, etc.). Within the visual feature scope, the features can be further classified as general features and domain-specific features. The former include color, texture and shape features while the latter is application dependent and may include, for example, human faces and fingerprints. Feature extraction algorithms of color-based, texture-based and shape-based image retrieval are available and in hot discussions. Here, we propose mixed content-based
image
retrieval,
combining
color-based,
texture-based and shape-based search. In our actual system, Fig.1. Classify the large image database into small distributed image
we utilize a mixed content-based image retrieval engine. Since
databases on categories
we aim at introducing the method and architecture of the
mixed query image retrieval system in this paper, we do not
receiving query request from management and control module,
intend to discuss content-based image retrieval techniques,
the system searches in the image database according to the
including multi-dimensional indexing techniques, in detail.
indexing tables and then return result to the management and
III.
ARCHITECTURE DESIGN OF MIXED QUERY IMAGE RETRIEVAL SYSTEM
control module. D.
There are four main parts in the mixed query image retrieval system. They are the user query interface, management
and
control
module,
text-based
and
content-based indexing database, and text and image
The database contains two kinds of data, text and image. The text database keep descriptive information related to images. The image database contains real pictures. The system architecture is shown in figure 3.
databases. A.
Database
IV.
User Query Interface
INTERACTIVE MODEL OF MIXED QUERY IMAGE RETRIEVAL SYSTEM
The user query interface should be friendly and flexible. To communicate with the user in a friendly manner, the query
Figure 4 shows the processing flow in the mixed query
interface is graphics-based. The interface collects the
image retrieval system. Several main parts in the process are
information needed from the users and displays the retrieval
explained briefly below.
results back to the users. Users can input keywords based on
¾
From information that the user input, extract keywords
category to assess to a certain kind of image database, and
based on category and according to these keywords,
then use the digital image characteristics, a picture or
select a corresponding clusters of indexing databases.
keywords of pictures to search in the image database. After
Use either content-based search or text-based search in
processing, the system will return results via the user query
the selected image databases.
interface. B.
Management and Control Module The module is the linchpin of the system, controlling and
managing the performance of the system. It processes user request, analyzes which type of the user request, text or image, process user request, and then return results to users. C.
Text-based indexing and image content-based indexing
Fig.3. System architecture of mixed query image retrieval system
Module There are two parts, that is, keywords based on category indexing and image content-based indexing. According to categories, a large image database is divided into
relatively
small
classified
image
databases.
Keyword-on-category indexing is established, associated with classified image databases. After processing text information in the management and control module, system automatically give control to the console of the corresponding image databases, according to the keyword-on-category indexing table. Image content-based indexing database keep image information and established image indexing tables. After
Fig.4. Interactive Model of Mixed query image retrieval system
¾
Gather candidate results from every selected indexing database, optimal resort, and select the most optimal set of results.
¾
According to the optimal set of image indexes, search in the image databases, and access to the actual image pictures.
¾
Organize the result pages, and return them to users. In the current design, process iii, process ii and process iv
occupy most of the internal processing time for the entire search process. In test queries, the processing time of process ii is less than 0.1 second. The processing time of process iii is less than 1 second and the time of process iv is also less than 1 second. The whole process can be completed within 2 seconds. In actual operation, especially in the real internet environment, the network transmission speed of process v can be improved by small size image output. V.
SYSTEM PROTOTYPE
VI.
CONCLUSION
This paper proposes a method to solve the problem of how to enhance the efficiency of image retrieval system applied to large scale image collections. Categorized segmentation of a large image database is the core concept. In this paper, we discuss a mixed query image retrieval system, combined both image content-based and text-bases search technologies, and present the system architecture that we design for this combined system and the system prototype. Experiments confirm that the mixed query image retrieval system can manage the large image database better and achieve high efficiency and better quality. Virtually, there are still many open issues needed to be solved before image retrieval systems can be put into practice. To achieve faster retrieval speed and make the image retrieval system scalable to large size image collections, multi-dimensional indexing technique is of great importance to the image retrieval system. In this
According to the system architecture design in the previous
paper we only discuss a kind of effective architecture of such
section, we have developed the prototype of mixed query
systems. In conclusion, integration of multiple technique and
image retrieval system. As shown below:
information sources of humans and computers will lead to a more successful image retrieval system. ACKNOWLEDGMENT I would like to thank Tony Chen for editing my English writing.
Fig.5a. System prototype
Fig.5b. System prototype
REFERENCES [1] Flickr web site, available at http://www.flickr.com
[10] [Coad 1991] P Coad and E Yourdon. “Object-Oriented Analysis”.
[2] IBM QBIC (Query By Image Content) home page, available at
Prentice-Hall, 1991.
http://wwwqbic.almaden.ibm.com/
[11] [Dijikstra 1979] E. Dijikstra. “Programming Considered as a Human
[3] Photobook system of MIT University home page, available at
Activity.” Classics in Software Engineering. Yourdon Press, 1979.
http://vismod.media.mit.edu/vismod/demos/photobook/index.html
[12] [Myers 1982] G. Myers. “Advances in Computer Architecture”. John
[4] VisualSEEK system of The DVMM Lab at Columbia University, available
Wiley and Sons, 1982.
at
[13] [Tsai 1988] J. Tsai and J. Ridge, “Intelligent Support for Specifications
http://www.ee.columbia.edu/ln/dvmm/researchProjects/MultimediaIndexing/
Transformation”. IEEE Software Vol. 5(6), p.34.
VisualSEEk/VisualSEEk.htm
[14] [Yourdon 1979] E. Yourdon and L. Constantine, “Structured Design”.
[5] LIKE.COM website, available at http://www.like.com
Prentice-Hall, Englewood-Cliffs, 1979.
[6] China Intelligence and Patent Website, available at http://www.cnipr.com/
[15] The google image search page, available at http://images.google.com
[7] Jeffrey Z. J. Zheng, Chris H. Zheng and Tosiyasu L. Kunii, “Concept Cell
[16] Thomas Deselasers, Tobias Weyand, Daniel Keysers, Wolfgang
Model for Knowledge Representation”, International Journal of Information
Ma-cherey and Hermann Ney, “FIRE in ImageCLEF 2005: Combining
Acquisition, Vol 1, No. 2, 149-168(2004)
Content-based Image Retrieval with Textual Information Retrieval”, In
[8 [Beynon-Davies 1993] P. Beynon-Davis. “Information System
Workshop of the Cross--Language Evaluation Forum (CLEF 2005), lecture
Development”. The Macmillan Press 1993.
Notes in Computer Science, volume 4022, pages 652-661, Vienna, Austria,
[9] [Booch 1990] G. Booch. “Object-oriented Analysis and Design”.
September 2005
Addison Wesley. 1990.