L. M. Patnaik

Image recommendation based on keyword relevance using absorbing Markov chain and image features

Image recommendation is an important feature of search engine, as tremendous amount of images are available online. It is necessary to retrieve relevant images to meet the user’s requirement. In this paper, we present an algorithm image recommendation with absorbing Markov chain (IRAbMC) to retrieve relevant images for a user’s input query. Images are ranked by calculating keyword relevance probability between annotated keywords from log and keywords of user input query. Keyword relevance is computed using absorbing Markov chain. Images are reranked using image visual features. Experimental results show that the IRAbMC algorithm outperforms Markovian semantic indexing (MSI) method with improved relevance score of retrieved ranked images.

DRSIG: Domain and Range Specific Index Generation for encrypted Cloud data

One of the most fundamental services of cloud computing is Cloud storage service. Huge amount of sensitive data is stored in the cloud for easy remote access and to reduce the cost of storage. The confidential data is encrypt before uploading to the cloud server in order to maintain privacy and security. All conventional searchable symmetric encryption(SSE) schemes enable the users to search on the entire index file. In this paper, we propose the Domain and Range Specific Index Generation(DRSIG) scheme that minimizes the Index Generation time. This scheme adopts collection sort technique to split the index file into D Domains and R Ranges. The Domain is based on the length of the keyword; the Range splits within the domain based on the first letter of the keyword. A mathematical model is used to encrypt the indexed keyword that eliminates the information leakage. The time complexity of the index generation is O(NT × 3) where NT - Number of rows in index document and 3 is Number of columns in index document. Experiments have been conducted on real world dataset to validate proposed DRSIG scheme. It is observed that DRSIG scheme is efficient and provide more secure data than Ranked Searchable Symmetric Encryption(RSSE) Scheme.

IR_URFS_VF: image recommendation with user relevance feedback session and visual features in vertical image search

In recent years, online shopping has grown exponentially and huge number of images are available online. Hence, it is necessary to recommend various product images to aid the user in effortless and efficient access to the desired products. In this paper, we present image recommendation framework with user relevance feedback session and visual features (IR_URFS_VF) to extract relevant images based on user inputs. User feedback is retrieved from image search history with clicked and un-clicked images. Image features are computed off-line and later used to find relevance between images. The relevance between images is determined by cosine similarity and are ranked based on clicked frequency and similarity score between images. Experiments results show that IR_URFS_VF outperforms CBIR method by providing more relevant ranked images to the user input query.

DAMS: Data Aggregation using Mobile Sink in Wireless Sensor Networks

WSNs consist of resource constrained sensor nodes that monitor the physical environment and transmit their data to the Sink through multi-hop communication. Mobile sinks are used to reduce the number of hops the data travels and thereby reducing the overall energy consumption. In this paper we propose Data Aggregation using Mobile Sink (DAMS) protocol that allows the mobile sink to collect data from WSNs where path of the mobile sink is not known apriori. The mobile sink halts at a point in the network and broadcasts an aggregate query. The average path length of a data packet is a constant in DAMS and hence it can withstand node failures. The performance analysis shows that DAMS incurs less energy consumption and improved packet delivery ratio in comparison to SinkTrail [1].

ACSIR: ANOVA Cosine Similarity Image Recommendation in vertical search

In today’s world, online shopping is very attractive and grown exponentially due to revolution in digitization. It is a crucial demand to provide recommendation for all the search engine to identify users’ need. In this paper, we have proposed a ANOVA Cosine Similarity Image Recommendation (ACSIR) framework for vertical image search where text and visual features are integrated to fill the semantic gap. Visual synonyms of each term are computed using ANOVA p value by considering image visual features on text-based search. Expanded queries are generated for user input query, and text-based search is performed to get the initial result set. Pair-wise image cosine similarity is computed for recommendation of images. Experiments are conducted on product images crawled from domain-specific site. Experiment results show that the ACSIR outperforms iLike method by providing more relevant products to the user input query.

RSSMSO Rapid Similarity Search on Metric Space Object Stored in Cloud Environment

This paper involves a cloud computing environment in which the dataowner outsource the similarity search service to a third party service provider. Privacy of the outsourced data is important because they may be confidential data. The data should be made available to the authorized client groups, but not to be revealed to the service provider in which the data is stored. Given this scenario, the paper presents a technique called RSSMSO which has build phase, query phase, data transformation and search phase. The build phase and the query phase are about uploading the data and querying the data respectively; the data transformation phase transforms the data before submitting it to the service provider for similarity queries on the transformed data; search phase involves searching similar object with respect to query object. The RSSMSO technique provides enhanced query accuracy with low communication cost. Experiments have been carried out on real data sets which exhibits that the proposed work is capable of providing privacy and achieving accuracy at a low cost in comparison with FDH

Two-Hop-Based Geographic Opportunistic Routing in WSNs

Existing work Geographic opportunistic routing (GOR) selects a forwarding sensor node to progress data packets on the basis of geographic distance. Similarly, the multipath routing uses multiple paths to achieve both reliability and delay. However, geographic opportunistic routing results in lower packet delivery rate and high latency. The multipath routing introduces channel contention, interference, and quick depletion of energy of the sensor node in an asymmetric link wireless environment. The existing work Efficient QoS-aware Geographic Opportunistic Routing (EQGOR) elects and prioritize the forwarding nodes to achieve different QoS parameters. However, in EQGOR, the count of forwarding nodes increases with the increase in the required reliability. To improve energy efficiency, delay, and successful ratio of packet delivery in WSNs, we propose a Two-Hop Geographic Opportunistic Routing (THGOR) protocol that selects a subset of 2-hop neighbors of node which has high packet reception ratio and residual energy at the next forwarder node, and the selected 1-hop neighbors of node has supreme coverage of 2-hop neighbors as relay nodes. THGOR is comprehensively evaluated through ns-2 simulator and compared with existing protocols EQGOR and GOR. Simulation results show that THGOR significant improvement in packet advancement, delay, reliable transmission, and energy efficient.

Split keyword fuzzy and synonym search over encrypted cloud data

A substitute solution for various organizations of data owners to store their data in the cloud using storage as a service(SaaS). The outsourced sensitive data is encrypted before uploading into the cloud to achieve data privacy. The encrypted data is search based on keywords and retrieve interested files by data user using a lot of traditional Search scheme. Existing search schemes supports exact keyword match or fuzzy keyword search, but synonym based multi-keyword search are not supported. In the real world scenario, cloud users may not know the exact keyword for searching and they might give synonym of the keyword as the input for search instead of exact or fuzzy keyword due to lack of appropriate knowledge of data. In this paper, we describe an efficient search approach for encrypted data called as Split Keyword Fuzzy and Synonym Search (SKFS). Multi-keyword ranked search with accurate keyword and Fuzzy search supports synonym queries are a major contribution of SKFS. The wildcard Technique is used to store the keywords securely within the index tree. Index tree helps to search faster, accurate and low storage cost. Extensive experimental results on real-time data sets shows, the proposed solution is effective and efficient for multi-keyword ranked search and synonym queries Fuzzy based search over encrypted cloud data.

FRORSS: Fast result object retrieval using similarity search on cloud

This paper involves a cloud computing environment in which the data owner out sources the similarity search service to a third party service provider. The user provides an example query to the server to retrieve similar data. Privacy of the outsourced data is important because they may be sensitive, valuable or confidential data. The data should be made available to the authorized client/client groups, but not to be revealed to the service provider in which the data is stored. Given this scenario, the paper presents a technique called FRORSS which has build phase, data transformation and search phase. The build phase is about uploading the data; the data transformation phase transforms the data before submitting it to the service provider for similarity queries on the transformed data; search phase involves searching similar object with respect to query object. Experiments have been carried out on real data sets which exhibits that the proposed work is capable of providing privacy and achieving accuracy at a lower value of result measure in comparision with FDH [1].