International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
Optimization of Image Search from Photo Sharing
Websites Using Personal Data
Mr. Naeem Naik
Prof. L.M R.J. Lobo
Walchand Institute of Technology, Solapur, India
Walchand Institute of Technology, Solapur India
Abstract
The present research aims at optimizing the image
search time of websites for the users by using their personal
data. Increasingly developed social sharing websites, like Flicker
and YouTube, allow the users to create, share, annotate and
comment Medias. This work describes a machine learningbased method for personalizing results of image search on
Flickr. Our method relies on metadata created by users through
their everyday activities on Flickr, namely the tags they used for
annotating their images and the groups to which they submitted
these images. This information captures user's tastes and
preferences in photography and can be used to personalize
image search results to the individual user. We validated our
approach by showing that it can be used to improve precision of
image search on Flickr for three ambiguous terms: “newborn,”
“tiger,” and “beetle.” In addition to improving search precision,
the tag-based approach can also be used to expand the search by
suggesting other relevant keywords (e.g., “pantheratigris,”
“bigcat” and “cub” for the query “tiger”).
Keywords- image search, optimization, metadata
1. INTRODUCTION
Web personalization refers to the process of
customizing Web experience to an individual user
(Mobasher, 2000). Personalization is used by online stores to
recommend relevant products to a particular user and to
customize a user’s shopping experience. It is used by
advertising firms to target ads to a particular user. Search
personalization has also been studied as a way to improve the
quality of Web search (Ma, 2007) by disambiguating query
terms based on user’s browsing history or by eliminating
irrelevant documents from search results.
Personalizing image search is an especially challenging
problem, because, unlike documents, images generally
contain little text that can be used for disambiguating terms.
Consider, for example, a user searching for photos of
“jaguars.”, Should the system return images of luxury cars or
spotted felines to the user? In this context, personalization
can help disambiguate query keywords used in image search
or to weed out irrelevant images from search results.
Therefore, if a user is interested in wildlife, the system will
show her images of the predatory cat of South America and
not of an automobile.
2. PROBLEM STATEMENT
In the Proposed System We propose a novel
personalized image search framework by simultaneously
considering user and query information. The user’s
preferences over images under certain query are estimated by
how probable he/she assigns the query-related tags to the
images.
 A ranking based tensor factorization model named
RMTF is proposed to predict user’s annotations to
the images.
 To better represent the query-tag relationship, we
propose to build user-specific topics and map the
queries as well as the users’ preferences onto the
learned topic spaces.
 User profile is proposed to be created.
 The proposed architecture is to be implemented
using three-tier architecture for more accuracy and
independency of layers.
 The comparison of two-tier and three-tier
architecture
2.1 PROPOSED METHODOLOGY
1. Principal Direction Divisive Partitioning (PDDP)
Algorithm:
Input: A n * m matrix (Documents Vers. Terms) c max =
Desired no. of clusters
Step 1: Initialize Binary Tree with a Root
Step 2: For c = 2, 3……. c max Do
Step 3: Select leaf node C with largest scatter value d, And L
& R : left & right children of C
Step 4: Compute vc = g(Mc) = ucT (Mc - wc eT)
Step 5: for i Є C, if vi <= 0, assign document I to L Else
assign document to R
Result: A binary tree with cmax leaf nodes forming partition
of a document set.
Time Complexity calculated: O(n2)
1.
2.
ISSN: 2231-5381
3. OBJECTIVES AND SCOPE
The objectives of the present research are
A Ranking based Multi-correlation Tensor Factorization
model is proposed to perform annotation prediction,
which is considered as users’ potential annotations for
the images.
We introduce User-specific Topic Modeling to map the
query relevance and user preference into the same user-
http://www.ijettjournal.org
Page 3601
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
3.
specific topic space. For performance evaluation, two
resources involved with users’ social activities are
employed. Experiments on a large scale Flickr dataset
demonstrate the effectiveness of the proposed method.
The Proposed Methodology reduces the Image search
time which in turn advantageous to the user.
4. DESIGN AND IMPLEMENTATION CONSTRAINTS
design view of a system.For the most Part this involves
modeling the vocabulary of the system, modeling
collaborations, or modeling schemas. Class diagrams are also
the foundation for a couple of related diagrams: Component
diagrams and Deployment diagrams. Class diagrams are
important not only for visualizing, specifying, and
documenting structural models, but also for constructing
executable systems through forward and reverse engineering.
The remarkable development of information on the
Web has forced new challenges for the construction of
effective search engines. The objective of this project is to
eliminate irrelevant search results by introducing Userspecific Topic Modeling to map the query relevance and user
preference into the same user-specific topic space. To better
represent the query-tag relationship, we build user-specific
topics and map the queries as well as the users’ preferences
onto the learned topic spaces.
The system provides proper search result by
reducing irrelevant searches. The proposed system has exact
intentions of the user queries and re-ranks the list results.
Given the large and growing importance of search engines,
personalized search has the potential to significantly improve
searching experience. It is very complicated for Web search
engines to satisfy the user information requirement only with
a short ambiguous query. To overcome such a basic difficulty
of information retrieval, personalized search, which is to
provide the customized search results to each user, is a very
promising solution. Fundamentally, in studying how a search
can be personalized, the most significant thing is to
accurately identify users’ information.
4.1 SYSTEM FEATURES
Two kinds of operations are handled in this
application. One is upload and share images and another is
search the image.
This application gives following details:
Pages containing images, each image have one or
more tags. Depending upon tags Clusters and Corpus of
images having same tags. Each page contains number of
images. Images are tagged depending upon category of image
e.g. fruit, car, etc. For each image one document of
classification of tags is created which is used for evaluating
output of Tensor Factorization Algorithm.
5. CLASS DIAGRAMS
Class diagrams are the most common diagram found in
modeling object-oriented system. A class diagram shows a
set of classes, interfaces and collaborations, and their
relationships. We use class diagrams to model the static
ISSN: 2231-5381
Figure 1- Class diagrams
6. OUTPUT SCREEN DESCRIPTION
The Output is best located on a web page given below in
Figure 7.1. The GUI for the application will look and seems
much user friendly. The Images that will be retrieved are
located in a systematic fashion. Other information related to
the results such as total no. of results retrieved and time to of
retrieval, shown in a simple but effective manner.
The description of links given:
1) Sign In: To sign in for user.
http://www.ijettjournal.org
Page 3602
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
2) Search Image(s): To display a web page containing
a textbox to input a query.
3) Upload An Image: To display a web page
containing a Browse button.
4) Tag An Image: To searchand tag a specific image.
5) Remove Image: To remove a uploaded image.
6) Create An Account: To create a new account for
new user.
7) Account Setting: To change the Account Settings
i.e (User Name and Passward).
8) Notes: To allow users to write their notes with
related to a search.
9) About The Project: To display the Authors
information, project information, details and
research module that were added to the project.
10) User’s query mapped to Specific Topic: To
display no. of topic generated. And their mapping to
user’s query
11) User’s Tag(s) to Image(s): To display the images
which are retrieved and also tagged by the user in
past.
Screen Shots
ISSN: 2231-5381
http://www.ijettjournal.org
Page 3603
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
7. CONCLUSION AND FINDINGS
In addition to creating content, users of Web 2.0
sites generate large quantities of metadata, or data about data,
that describe their interests, tastes and preferences. These
metadata, in the form of tags and social networks, are created
mainly to help users organize and manage their own content.
These types of metadata can also be used to target relevant
content to the user through recommendation or
personalization.
This proposed work describes a machine learning-based
method for personalizing results of image search on Flickr.
Our method relies on metadata created by users through their
everyday activities on Flickr, namely the tags they used for
annotating their images and the groups to which they
submitted these images. This information captures user's
tastes and preferences in photography and can be used to
personalize image search results to the individual user. We
validated our approach by showing that it can be used to
improve precision of image search on Flickr for three
ambiguous terms: “newborn,” “tiger,” and “beetle.” In
ISSN: 2231-5381
http://www.ijettjournal.org
Page 3604
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
addition to improving search precision, the tag-based
approach can also be used to expand the search by suggesting
other relevant keywords (e.g., “pantheratigris,” “bigcat” and
“cub” for the query “tiger”).
Comparison of Two-Tier (Blue) and Three-Tier( Red)
architecture for no. Of results retrieved:
We fired 10 same queries against Two-Tier and Three-Tier
architecture to compare no. Of results retrieved.
We found that out of these 10 queries, 7 times the no. Of
results retrieved with Three-Tier architecture were more than
Two-Tier architecture.
Performance values:
Two-Tier architecture
Three-Tier architecture
3 times / 10
7 times / 10
= 30 %
= 70 %
 Jin, R., Si, L., & Zhai, C. (2006) A study of mixture
models for collaborative filtering. Information
Retrieval 9(3):357–382.
 J. Tang, H. Li, G. Qi and T. Chua, “ Image
Annotation by Graph-Based Inference With
Integrated
Multiple/Single
Instance
Representations,” in IEEE Trans. Multimedia, 2010,
vol. 12, no. 2, pp. 131–141, 2010.
 M. J. Carman, M. Baillie, and F. Crestani, “Tag data
and personalized information retrieval,” in SSM,
2008, pp. 27–34.
 R. J¨aschke, L. B. Marinho, A. Hotho, L. SchmidtThieme, and G. Stumme, “Tag recommendations in
folksonomies,” in PKDD, 2007, pp. 506–514.
 R. J¨aschke, L. B. Marinho, A. Hotho, L. SchmidtThieme, and G. Stumme, “Tag recommendations in
social bookmarking systems,” AI Commun., vol. 21,
no. 4, pp. 231–247, 2008
 P. Symeonidis, A. Nanopoulos and Y.
Manolopoulos, “A Unified Framework for
Providing Recommendations in Social Tagging
Systems Based on Ternary Semantic Analysis,”
IEEE Trans. Knowl. Data Eng., vol. 22,no. 2, pp.
179–192, 2010.
References
 Agrawal, R., & Srikant, R. (1994). Fast algorithms
for mining association rules. In Bocca, J. B., Jarke,
M.& Zaniolo, C. (Eds.), Proceedings of the 20th Int.
Conf. Very Large Data Bases, VLDB (pp. 487—
499). Morgan Kaufmann.
 Breese, J., Heckerman, D.& Kadie, C. (1998).
Empirical analysis of predictive algorithms for
collaborative filtering. In Proceedings of the 14th
Annual Conference on Uncertainty in Artificial
Intelligence (pp. 43—52). San Francisco, CA:
Morgan Kaufmann.
 Dempster, A. P., Laird, N.M. & Rubin, D.B. (1977).
Maximum likelihood from incomplete data via the
em algorithm. Journal of the Royal Statistical
Society. Series B (Methodological) 39(1), 1-38.
 G. Zhu, S. Yan, and Y. Ma, “Image Tag Refinement
Towards Low-Rank, Content-Tag Prior and Error
Sparsity,” in ACM Multimedia, 2010, pp. 461–470.
 Golder, S.A. & Huberman, B.A.(2006). The
structure of collaborative tagging systems. Journal
of Information Science 32(2), 198-208.
ISSN: 2231-5381
http://www.ijettjournal.org
Page 3605