International Journal of Emerging Computer Science & Technology
IMAGE RE-RANKING BASED ON TOPIC DIVERSITY
S.RIZWANA BANU,
RESEARCH SCHOLAR,
DEPARTMENT OF COMPUTER SCIENCE,
RAJAH SERFOJI GOVT COLLEGE,
THANJAVUR 613005, Tamilnadu -India
Abstract— Social media sharing websites allow users to
annotate images with free tags, which significantly
contribute to the development of the web image retrieval.
Tag-based image search is an important method to find
images shared by users in social networks. However, how
to make the top ranked result relevant and with diversity
is challenging. In this paper, we propose a topic diverse
ranking approach for tag-based image retrieval with the
consideration of promoting the topic coverage
performance. First, we construct a tag graph based on the
similarity between each tag. Then community detection
method is conducted to mine the topic community of each
tag. After that, inter-community and intra community
ranking are introduced to obtain the final retrieved
results. In the inter-community ranking process, an
adaptive random walk model is employed to rank the
community based on the multi-information of each topic
community. Besides, we build an inverted index structure
for images to accelerate the searching process.
Experimental results on Flickr dataset and NUS-Wide
datasets show the effectiveness of the proposed approach.
Keywords— Tag-based image, inter-community, intra
community.
I. INTRODUCTION
In the era of big data, the amount of information made
available to users far exceeds their capacity to discover and
understand it. For instance, a user on Twitter may receive an
overwhelming volume of notifications if her message is
retreated by too many people in a short period. Moreover, the
timeliness of information filtering and delivery is of great
ISSN NO: 0974-2358 Volume 5(Aug 2018) Pages(10-18)
Dr.M.Chidambaram,
ASSISTANT PROFESSOR,
DEPARTMENT OF COMPUTER SCIENCE,
RAJAH SERFOJI GOVT COLLEGE,
THANJAVUR 613005, Tamilnadu -India.
importance. For example, a user would like to receive instant
updates of the hottest topics on social news and entertainment
websites (e.g., on reddit.com). Thus, the efficient filtering and
monitoring of rapid streams is key to many emerging
applications. "Most of the concordances produced in recent
times and with the aid of computer software use both the
KWIC (keyword in context) and KWICn (keyword in centre)
formats, which lists the keyword, usually highlighted in bold
text in a consistent position, within a limited amount of
context text, i.e. three [or] four words of the text prior to the
keyword and the same amount of text following. This format
is extremely useful in that the keyword is easily identified
together with its context. The Concordance of the Roman
Missal is produced in both the KWIC and KWICn formats and
is noteworthy in that each word form is listed as it appears in
the text, that is, it is un-lemmatized." With the development of
social media based on web 2.0, amounts of images and videos
spring up everywhere on the internet. This phenomenon has
brought great challenges to multimedia storage, indexing and
retrieval. Generally speaking, tag-based image search is more
commonly used in social media than content based image
retrieval and content understanding [2-10, 13, 16, 22, 32, 38,
40,44, 46, 47-51,64,65]. Thanks to the low relevance and
diversity performance of initial retrieval results, the ranking
problem in the tag-based image retrieval has gained
researchers’ wide attention [42, 43, 47-50].
Nonetheless, the following challenges block the path for the
development of re-ranking technologies in the tag-based
image retrieval.
1) Tag mismatch. Social tagging requires users to label their
uploaded images with their own keywords and share with
International Journal of Emerging Computer Science & Technology

others [25]. Different from ontology based image annotation; hierarchical [56]. They first organize images to different leaf
there is no predefined ontology or taxonomy in social image topic, and then define the topic cover score based on topic list,
tagging. Every user has its own habit to tag images. Even for and finally use a greedy algorithm to obtain the highest topic
the same image, tags contributed by different users will be of cover score list. Dang-nguyen et al. [57] first propose a
great difference [25, 43, 47]. Thus, the same image can be clustering algorithm to obtain a topic tree, and then sort topics
interpreted in several ways with several different tags according to the number of images in the topic. In each
according to the background behind the image. In this case, cluster, the image uploaded by the user who has highest visual
many seemingly irrelevant tags are introduced. score is selected as the top ranked image. The second image is
the one which has the largest distance to the first image. The
2) Query ambiguity. Users cannot precisely describe their
third image is chosen as the Image with the largest distance to
request with a single word and tag suggestion systems always
both two previous images, and so on. In our previous work
recommend words that are highly correlated to the existing tag
[47], the diversity is achieved based on social user re-ranking.
set. Besides, polysemy and synonyms are the other causes of
We regard the images uploaded by the same user as a cluster
the query ambiguity.
and we pick one image from each cluster to achieve the
diversity. Most papers consider the diversity from visual
Thus, a fundamental issue in the ranking of the tag-based
perspective and achieve it by applying clustering on visual
social image retrieval is how to solve these problems reliably.
features [47, 55-59]. In this paper, we focus on the topic
As far as the “tag mismatch” problem is concerned, tag
diversity. We first group all the tags in the initial retrieval
refinement [1, 20, 24-26, 45], tag relevance ranking [17, 35,
image list to make the tags with similar semantic be the same
45] and image relevance ranking [3, 7, 15, 21, 27, 33, 34]
cluster, then assign images into different clusters. The images
approaches have been dedicated to overcome it. As for the
within the same cluster are viewed as the ones with similar
“query ambiguity” problem, an effective approach is to
semantics. After ranking the clusters and images in each
provide diverse retrieval results that cover multiple topics
cluster, we select one image from each cluster to achieving our
underlying a query. Currently, image clustering [10, 43] and
semantic diversity.
duplicate removal [5-6, 9, 28, 29-31] are the major approaches
in settling the diversity problem. However, most of the
II. LITERATURE SURVEY
literature regards the diversity problem as to promote the
Study on Various Techniques of Image Enhancement
visual diversity performance, but the promotion of the
semantic coverage is often ignored. To diversify the top
Social media sharing websites allow users to annotate images
ranked search results from the semantic aspect, the topic
with free tags, which significantly contribute to the
community belongs to each image should be considered.
development of the web image retrieval.Tag-based image
search is an important method to find images shared by users
In recent years, more and more scholars pay attention to
in social networks. However, how to make the top ranked
retrieval result’s diversity [47, 51,55-63]. In [55], the authors
result relevant and with diversity is challenging. In this paper,
first apply graph clustering to assign the images to clusters,
a topic diverse ranking approach for tag-based image retrieval
and then utilize random walk to obtain the final result. The
with the consideration of promoting the topic coverage
diversity is achieved by set the transition probability of two
performance. First, we construct a tag graph based on the
images in different clusters higher than that in the same
similarity between each tag. Then community detection
cluster. Tian et al. think the topic structure in the initial list is
International Journal of Emerging Computer Science & Technology
method is conducted to mine the topic community of each tag.
After that, inter-community and intra-community ranking are
introduced to obtain the final retrieved results. In the inter-
community ranking process, an adaptive random walk model
is employed to rank the community based on the multi-
information of each topic community. Besides, we build an
inverted index structure for images to accelerate the searching
process. Experimental results on Flickr dataset and NUS-Wide
datasets show the effectiveness of the proposed approach.
2. Exploiting Query Reformulations for Web Search
Result Diversification
When a Web user’s underlying information need is not clearly
specified from the initial query, an effective approach is to
diversify the results retrieved for this query. In this paper, we
introduce a novel probabilistic framework for Web search
result diversification, which explicitly accounts for the various
aspects associated to an underspecified query. In particular,
we diversify a document ranking by estimating how well a
given document satisfies each uncovered aspect and the extent
to which different aspects are satisfied by the ranking as a
whole. We thoroughly evaluate our framework in the context
of the diversity task of the TREC 2009 Web track. Moreover,
we exploit query reformulations provided by three major Web
search engines (WSEs) as a means to uncover different query
aspects. The results attest the effectiveness of our framework
when compared to state-of-the-art diversification approaches
in the literature. Additionally, by simulating an upper-bound
query reformulation mechanism from official TREC data, we
draw useful insights regarding the effectiveness of the query
reformulations generated by the different WSEs in promoting
diversity. Categories and Subject Descriptors
3. Tag-Based Social Image Search with Visual-Text Joint
Hypergraph Learning
Tag-based social image search has attracted great interest and
how to order the search results based on relevance level is a
research problem. Visual content of images and tags have both
been investigated. However, existing methods usually employ
tags and visual content separately or sequentially to learn the
image relevance. This paper proposes a tag-based image
search with visual-text joint hypergraph learning. We
simultaneously investigate the bag-of-words and bag-of-
visual-words representations of images and accomplish the
relevance estimation with a hypergraph learning approach.
Each textual or visual word generates a hyperedge in the
constructed hypergraph. We conduct experiments with a real-
world data set and experimental results demonstrate the
effectiveness of our approach.
4. A General Framework for Edited Video and Raw Video
Summarization
Overall, our work can be divided into three folds: 1) Four
models are designed to capture the properties of video
summaries, i.e., containing important people and objects
(importance), representative to the video content
(representativeness), no similar key-shots (diversity) and
smoothness of the storyline (storyness). Specifically, these
models are applicable to both edited videos and raw videos. 2)
A comprehensive score function is built with the weighted
combination of the aforementioned four models. Note that the
weights of the four models in the score function, denoted as
property-weight, are learned in a supervised manner. Besides,
the property-weights are learned for edited videos and raw
videos, respectively. 3) The training set is constructed with
both edited videos and raw videos in order to make up the lack
of training data. Particularly, each training video is equipped
with a pair of mixing-coefficients which can reduce the
structure mess in the training set caused by the rough mixture.
5. for Your Eyes Only: Learning to Summarize First-
Person Videos
With the increasing amount of video data, it is desirable to
highlight or summarize the videos of interest for viewing,
search, or storage purposes. However, existing summarization
approaches are typically trained from third-person videos,
International Journal of Emerging Computer Science & Technology
which cannot generalize to highlight the first-person ones. By
advancing deep learning techniques, we propose a unique
network architecture for transferring spatiotemporal
information across video domains, which jointly solves
metric-learning based feature embedding and keyframe
selection via Bidirectional Long Short-Term Memory
(BiLSTM). A practical semi-supervised learning setting is
considered, i.e., only fully annotated third-person videos,
unlabeled first-person videos, and a small amount of annotated
first-person ones are required to train our proposed model.
Qualitative and quantitative evaluations are performed in our
experiments, which confirm that our model performs
favorably against baseline and state-of-the-art approaches on
first-person video summarization.
6. Automatic Meaning Discovery Using Google
A method to automatically extract the meaning of words and
phrases from the world-wide-web using Google page counts.
The approach is novel in its unrestricted problem domain,
simplicity of implementation, and manifestly ontological
underpinnings. The world-wide-web is the largest database on
earth, and the latent semantic context information entered by
millions of independent users averages out to provide
automatic meaning of useful quality. We demonstrate positive
correlations, evidencing an underlying semantic structure, in
both numerical symbol notations and number-name words in a
variety of natural languages and contexts. Next, we
demonstrate the ability to distinguish between colors and
numbers, and to distinguish between 17th century Dutch
painters; the ability to understand electrical terms, religious
terms, and emergency incidents; we conduct a massive
experiment in understanding WordNet categories; and finally
we demonstrate the ability to do a simple automatic English-
Spanish translation.
Objects can be given literally, like the literal four-letter
genome of a mouse, or the literal text of War and Peace by
Tolstoy. For simplicity we take it that all meaning of the
object is represented by the literal object itself. Objects can
also be given by name, like “the four-letter genome of a
mouse,” or “the text of War and Peace by Tolstoy.” There are
also objects that cannot be given literally, but only by name,
and that acquire their meaning from their contexts in
background common knowledge in humankind, like “home”
or “red.” To make computers more intelligent one would like
to represent meaning in computer-digestable form. Long-term
and labor-intensive efforts like the Cyc project [18] and the
WordNet project [29] try to establish semantic relations
between common objects, or, more precisely, names for those
objects. The idea is to create a semantic web of such vast
proportions that rudimentary intelligence, and knowledge
about the real world, spontaneously emerge. This comes at the
great cost of designing structures capable of manipulating
knowledge, and entering high quality contents in these
structures by knowledgeable human experts. While the efforts
are long-running and large scale, the overall information
entered is minute compared to what is available on the world-
wide-web. The rise of the world-wide-web has enticed
millions of users to type in trillions of characters to create
billions of web pages of on average low quality contents. The
sheer mass of the information available about almost every
conceivable topic makes it likely that extremes will cancel and
the majority or average is meaningful in a low-quality
approximate sense. We devise a general method to tap the
amorphous low-grade knowledge available for free on the
world-wide-web, typed in by local users aiming at personal
gratification of diverse objectives, and yet globally achieving
what is effectively the largest semantic electronic database in
the world. Moreover, this database is available for all by using
any search engine that can return aggregate page-count
estimates for a large range of search-queries, like Google.
7. The Use of MMR, Diversity-Based Reranking for
Reordering Documents and Producing Summaries
This paper presents a method for combining query-relevance
with information-novelty in the context of text retrieval and
summarization. The Maximal Marginal Relevance (MMR)
criterion strives to reduce redundancy while maintaining query
International Journal of Emerging Computer Science & Technology
relevance in re-ranking retrieved documents and in selecting
apprw priate passages for text summarization. Preliminary
results indicate some benefits for MMR diversity ranking in
document retrieval and in single document summarization.
The latter are borne out by the recent results of the SUMMAC
conference in the evaluation of summarization systems.
However, the clearest advantage is demonstrated in
constructing non-redundant multi-document summaries, where
MMR results are clearly superior to non-MMR passage
selection.
With the continuing growth of online information, it has
become increasingly important to provide improved
mechanisms to find information quickly. Conventional IR
systems rank and assimilate documents based on maximizing
relevance to the user query [l, 51. In cases where relevant
documents are few, or cases where very-high recall is
necessary, pure relevance ranking is very appropriate. But in
cases where there is a vast sea of potentially relevant
documents, highly redundant with each other or (in the
extreme) containing partially or fully duplicative information
we must utilize means beyond pure relevance for document
ranking. A new document ranking method is one where each
document in the ranked list is selected according to a
combined criterion of query relevance and novelty of
information. The latter measures the degree of dissimilarity
between the document being considered and previously
selected ones already in the ranked list. Of course, some users
may prefer to drill down on a narrow topic and others a
panoramic sampling bearing relevance to the query. Best is a
user-tunable method; Maximal Marginal Relevance (MMR)
provides precisely such functionality, as discussed below.
8. Social Re-ranking using Tag Based Image Search
The growing amount of user-tagged multimedia has led social
image analysis and retrieval gain importance which has helped
people organize and access user tagged multimedia. User
tagging is uncontrolled, involves ambiguity and highly
personalized hence a fundamental question arises how to
interpret the relevance of user-contributed tag with respect to
the visual content described by the tag. In this work, image’s
relevance and diversity is considered and a social reranking
system for tag-based image retrieval is proposed. According to
respective visual information, semantic information and social
clues the images are re-ranked. The users that have higher
contribution to the given query are ranked higher. Then
sequential checking time stamp ranking is performed in which
the desired output is obtained on basis of title information and
the recent time stamp which enhances the diversity
performance of image ranking system. It also counts number
of views utilized to improve the relevance performance of the
image retrieval results. The final retrieved results are
composed of the selected images. A keyword relevancy match
the data is retrieved for the social image dataset to accelerate
the searching process.
9. Tag Image Re-Ranking for Social Dataset
Tag-based image search is an important method to process
images contributed by social users in social media sharing
websites like Flickr. Since user tagging is known to be
uncontrolled, ambiguous, and overly personalized, a
fundamental problem is how to interpret the relevance of a
user contributed tag with respect to the visual content the tag
is describing. However, existing ranking methods for tag-
based image search frequently return results that are irrelevant,
low-diversity or time-consuming. In this work, it proposes a
social re-ranking system for tag-based image retrieval with the
consideration of image’s relevance and diversity. We aim at
reranking images according to their visual information,
semantic information and social clues. The initial results
include images contributed by different social users. Usually
each user contributes several images. First we sort these
images by interuser re-ranking. Users that have higher
contribution to the given query rank higher. Then we
sequentially check title and time stamp ranking in which the
desired output will get on the basis of title information and the
recent time stamp which enhance the diversity performance of
image ranking system also it count number of views has been
utilized to improve the relevance performance of the image
International Journal of Emerging Computer Science & Technology

retrieval results. These selected images compose the final
retrieved results. We build an identify keyword relevancy
match the data is retrieved for the social image dataset to
accelerate the searching process. Experimental results on
Flickr dataset show that our tag image re-ranking method is
effective and efficient.
10. IMAGE RETRIEVAL AND RE-RANKING
TECHNIQUES - A SURVEY
There is a huge amount of research work focusing on the
searching, retrieval and re-ranking of images in the image
database. The diverse and scattered work in this domain needs
to be collected and organized for easy and quick reference.
Relating to the above context, this paper gives a brief
overview of various image retrieval and re-ranking techniques.
The paper proceeds through the core architecture of image
harvesting and retrieval system to the different Re-ranking
techniques.
Image retrieval is a key issue of user concern. Normal
way of image retrieval is the text based image retrieval
technique (TBIR)[12]. TBIR-needs rich semantic textual
description of web images .This technique is popular but
needs very specific description of the query which is tedious
and not always possible. Therefore generally the process of
diversity and relevance performance. The tag graph
construction based on each tag’s word vector and community
mining approach are employed in our approach to detect topic
community. We rank each mined community according to
their relevance level to the query. In the inter-community
ranking process, an adaptive random walk model is employed
to accomplish the ranking based on the relevance of each
community with respect to the query, pair-wise similarity
between each community, and the image number in each
community. With the adaptive random walk model, the
community that possesses the bigger semantic relevance value
with the query and larger confidence value will be ranked
higher. It computes the similarity between the user oriented
image set and query based on the co-occurrence tag
mechanism. Sort the image collections of different users
according to their contributions, (i.e.) the number of co-
occurrence tags of query in users’ tag sets.
IV. METHODOLOGY
The proposed social re-ranking system includes two main
sections: online and offline as shown in Fig.1.
USER ADMIN
ALL TYPE SEARCH

image search includes searching of image based on keyword CLIENT

UPLOADING
typed. The process that occurs in the background is not so IMAGES AUTHENTICATION
simple though. When query is entered in the search box for INTERN SEARCH

searching the image, it is forwarded to the server that is

connected to the internet. The server gets the URL’s of the
INTRA SEARCH
images based on the tagging of the textual word from the GRAPHICAL APPROACH
TO ALL IMAGES AND
REQUIREMENYTS INFORMATION ABOUT
SERVER
internet and sends them back to the client. THE IMAGES

III. PROPOSED SYSTEM

Fig.1. social re-ranking system.
Our proposed work is to construct the tag graph and mine the
1) Inverted index structure construction for image
topic community to diversify the semantic information of the
dataset is built to accelerate the retrieval speed. 2) Feature
retrieval results. We propose a topic diverse ranking approach
extraction. In this paper, the visual feature, semantic feature,
considering the topic coverage of the retrieved images. The
views and Timestamps for the images dataset are extracted.
intercommunity ranking method and intra-community ranking
Semantic feature refers to the co occurrence word set of query
methods are proposed to achieve a good trade-off between the
tags and the tags of the images. The online parts consist of the
International Journal of Emerging Computer Science & Technology
following three steps: 1) Keyword matching. For an input
query, the system will return the initially retrieved results by
keyword matching.
The following two online steps are all conducted to
re-rank the initial results. 2) Inter-user re-ranking. It is applied
to rank the corresponding users with the considering their
contributions to the given query. 3) Intra-user re-ranking. To
determine the relevance level of each image by fusing the
visual, semantic and views information into a unified system a
regularization framework is proposed. The most relevant
image in each ranked user’s image set is sequentially selected.
These selected images constitute our re-ranking results.
Different approaches can be used in tag based image
search as follows [14]: a) VR: View-based re-ranking, it is a
measure that ranks the initial results by views in a descending
order. b) VUR: View and user based re-ranking. This
approach is based on VR. In this approach the images which
share same user are removed and the final re-ranked results
are obtained. The image with the largest views for a user is
kept in the top ranked results. c) SR: Social re-ranking
promotes the relevance and diversity performance of our
results. Diversity performance is boosted by using User
information. A regularization framework that combines
semantic, visual and views information is put forward to
improve the relevance performance. d) TSR: The method
proposed of Time stamp information is used to search tag
based images by considering the time stamp information i.e.
fused with Visual features, semantic Features, Views and time
required to retrieve image from database. Time consumption
in searching the result is reduced and desired output is
obtained.
V. RESULT AND DISCUSSION
Re-ranking system for tag based images in social dataset is
proposed. The contributions can be summarized as follows: A
tag based images search for social dataset is proposed. Firstly
the input query is taken from user. A particular meaningful
keyword considers for example “animal” is taken. Then it will
match the keyword. The social re-ranking system includes
online and offline as two main sections. Tag image dataset is
used in offline section. Keyword matching and image re-
ranking are done through offline mode. Online section uses
the tag image dataset in offline mode.
Keyword is identified after this and relevancy matching is
done. The synonyms are taken or identified for given query
i.e. for example synonyms of “animals”. Keyword matching
helps to identify keyword relevancy and the matched data is
retrieved. The retrieved data is gone through three steps:
1) Inter-User Ranking by Query – Users images are ranked
according to given query. A good trade-off is achieved
between the diversity and relevance performance with this
ranking system. This effectively eliminates similar images
from the same user in a ranked result.
2) Time Stamp Ranking – The result of inter user ranking is
gone through title and time stamp ranking. In this the desired
output will is obtained on the basis of title information and the
recent time stamp. This enhances the diversity performance of
image ranking system.
3) Views Ranking – In social media community the views of
an image is an important feature. It indicates the click count of
image. The click count is used to improve the relevance
performance of image retrieval results.
A. EXPERIMENTAL RESULT:
After all these process the system is ranked and the desired
image is obtained. Experimental result dataset shows that
social re-ranking method is effective and efficient. As Shown
in Table.1. Social re-ranking time for retrieval has improved.
VR- View based Re-ranking
VUR- View and User based Re-ranking
RR- Relevance Re-ranking
CRR - Co-occurrence based Re-ranking
DRR - Diverse Relevance Re-ranking
SR - Social Re-ranking without Timestamps
SR - Social Re-ranking with Timestamps
International Journal of Emerging Computer Science & Technology

[5] A. Lacerda and N. Ziviani, “Building user profiles to

3500
improve user experience in recommender systems,” in
3000 WSDM, 2013, pp. 759–764.
2500 [6] M. Busch, K. Gade, B. Larson, P. Lok, S. Luckenbill, and
J. J. Lin, “Earlybird: Real-time search at twitter,” in ICDE,
2000
Series 1 2012, pp. 1360– 1369.
1500 Series 2 [7] L. Wu, W. Lin, X. Xiao, and Y. Xu, “LSII: an indexing
1000 structure for exact real-time search on microblogs,” in ICDE,
2013, pp. 482–493.
500
[8] J. Zobel and A. Moffat, “Inverted files for text search
0 engines,” ACM Comput. Surv., vol. 38, no. 2, 2006.
VR VUR RR CRR DRR SR SR
[9] R. Fagin, A. Lotem, and M. Naor, “Optimal aggregation

VI. CONCLUSION algorithms for middleware,” J. Comput. Syst. Sci., vol. 66, no.
In this paper, we propose a topic diverse re-ranking method 4, pp. 614–656, 2003.
for tag-based image retrieval. In this topic diverse re-ranking [10] A. Z. Broder, D. Carmel, M. Herscovici, A. Soffer, and J.
method, inter-community ranking and intra-community Y. Zien, “Efficient query evaluation using a two-level retrieval
ranking are carried out to get satisfactory retrieved results. Tag process.” in CIKM, 2003, pp. 426–434.
graph construction and community detection are two effective [11] R. Cilibrasi, and P. Vitanyi, “The Google Similarity
ways to enhance the diversity. Besides, each tag’s word vector Distance”. IEEE Trans. Knowledge and Data Engineering,
is trained by using the Word2vec model based on the English 19(3):1065-1076, 2007.
Wikipedia corpus to enhance the relevance performance of the [12] X. Qian, H. Wang, G. Liu, and X. Hou, “HWVP:
retrieved results. Hierarchical Wavelet Packet Texture Descriptors and Their
Applications in Scene Categorization and Semantic Concept

References Retrieval”. Multimedia Tools and Applications, May 2012.

[1] P. Haghani, S. Michel, and K. Aberer, “The gist of
everything new: personalized top-k processing over web 2.0
streams.” in CIKM, 2010, pp. 489–498.
[2] K. Mouratidis and H. Pang, “Efficient evaluation of
continuous text search queries,” IEEE Trans. Knowl. Data
Eng., vol. 23, no. 10, pp. 1469–1482, 2011.
[3] N. Vouzoukidou, B. Amann, and V. Christophides,
“Processing continuous text queries featuring
homogeneous scoring functions.” in CIKM, 2012, pp. 1065–
1074.
[4] A. Hoppe, “Automatic ontology-based user profile
learning from heterogeneous web resources in a big data
context.” PVLDB, pp. 1428–1433, 2013.
[13] X. Lu, Y. Yuan, X. Zheng, Jointly Dictionary Learning
for Change Detection in Multispectral Imagery, IEEE Trans.
Cybernetics, vol. 47, no. 4, pp. 884-897, 2017.
[14] J. Carbonell, and J. Goldstein, “The use of MMR,
diversity based re-ranking for reordering documents and
producing summaries”. SIGIR 1998.
[15] Wu, J. Wu, and M. Lu, “A Two-Step Similarity Ranking
non- Scheme for Image Retrieval. In Parallel Architectures”.
Algorithms and Programming, pp. 191-196, IEEE, 2014.
[16] G. Ding, Y. Guo, J. Zhou, et al., Large-Scale Cross-
Modality Search via Collective Matrix Factorization Hashing.
IEEE Transactions on Image Processing, 2016, 25(11): 5427-
5440.
International Journal of Emerging Computer Science & Technology

[17] G. Agrawal, and R. Chaudhary, “Relevancy tag ranking”.

In Computer and Communication Technology, pp. 169-173,
IEEE, 2011.
[18] L. Chen, S. Zhu, and Z. Li, “Image retrieval via improved
relevance ranking”. In Control Conference, pp. 4620-4625,
IEEE, 2014.
[19] L. Wu, and R. Jin, “Tag completion for image retrieval”.
Pattern Analysis and Machine Intelligence, IEEE Transactions
on, 35(3), 716-727, 2013.
[20] Y. Yang, Y. Gao, H. Zhang, and J. Shao, “Image Tagging
with Social Assistance”. ICMR, 2014.
[21] L. Chen, D. Xua, and I. Tsang, “Tag-based image
retrieval improved by augmented features and group-based
refinement”. Multimedia, IEEE Transactions on, 14(4), 1057-
1067, 2012.
[22] Z. Lin Z, G. Ding, J. Han, et al., Cross-View Retrieval via
Probability-Based Semantics-Preserving Hashing, IEEE
Transactions on Cybernetics ,vol. PP, no.99, pp.1-14 doi:
10.1109/TCYB.2016.2608906.
[23] R. Agrawal, S. Gollapudi, A. Halverson, and S. Ieong,
“Diversifying search results”. In WSDM, pages 5–14, 2009.