sonali bodkhe - Academia.edu (original) (raw)

Papers by sonali bodkhe

This paper provides the information about image retrieval using the concept of Attribute Base Ima... more This paper provides the information about image retrieval using the concept of Attribute Base Image Retrieval (ABIR) and Content Base Image Retrieval (CBIR). The retrieval method nominate in present paper utilizes the fusion of the images multimodal information (visual and textual) which is a recent trend in image retrieval researches. Image retrieval is the science of locating images from a large database or image sequences that fulfill the specified image need. An image retrieval technique is based on the Scale-Invariant Feature Transform (SIFT) method is present in this project. It joins two different data mining techniques to get semantically related images they are as: association and clustering rules mining algorithm which is nothing but the Bag of Features (BoF). BoF methods are based on order less collection of quantized local image descriptors they remove spatial information and are therefore computationally and conceptually simpler than many alternative methods.

This paper provides the information about an effective method for MRI brain image enhancement. Th... more This paper provides the information about an effective method for MRI brain image enhancement. The automatic segmentation of brain-tissue has led to the variation in the images due to different scanning and the imaging protocols which makes the image unclear and thus application is hampered. The transfer learning with weighted SVM enables training data to minimize classification errors as the classification scheme needs only a small amount of representative data. Therefore a new optimally standardized method is presented for scanned image segmentation using Transfer Learning with Weighted Support Vector Machine and then further improving the training data quality by Vector Sparse Representation using Iterative Algorithm for Quaternion Matrix Analysis over Reflexive Matrices.

2008 International Conference on Computer Science and Information Technology, 2008

... Segmentation of MRI images using intensity values is severely limited owing to in-homogeneiti... more ... Segmentation of MRI images using intensity values is severely limited owing to in-homogeneities and partial volume effects. Edge based segmentation methods suffer from unsharp edges and gaps in boundaries. A number of other methods are also used. ...

Ad hoc wireless networks are defined as the category of wireless networks that utilizes multi-hop... more Ad hoc wireless networks are defined as the category of wireless networks that utilizes multi-hop radio relaying and are capable of operating without the support of any fixed infrastructure hence, they are called infrastructure less networks. The lack of any central coordination makes them more vulnerable to attacks than wired networks. Due to some unique characteristics of MANETs, prevention methods alone are not sufficient to make them secure therefore, detection should be added as another defense before an attacker can breach the system. Network intrusion detection is the process of monitoring the events occurring in the network and analyzing them for signs of intrusions, defined as attempts to compromise the confidentiality. In this paper, we define and discuss various techniques of Intrusion Detection. We also present a description of routing protocols and types of security attacks possible in the network.

Duplicate detection is the process of identifying multiple representations of same real world ent... more Duplicate detection is the process of identifying multiple representations of same real world entities. The proposed System will compare Duplication Detection Method for best results. These methods are used for removing duplicate data and to cut redundancy. The first is based on Novel progressive duplicate detection algorithms that will significantly increase the efficiency of finding duplicates if the execution time is limited. This method maximizes the gain of overall process within the time available by reporting most results earlier than traditional approaches. The second is based on Secure Hashing Algorithm which will detect duplicate data for performing data de-duplication task to overcome the issues of time and to cut hash collision. This architecture will be useful for storage server where a huge amount of data is stored every day and software industries always looks for new developments so that they can keep their storage systems up to date and free for efficient use of it.

The present paper provides the information about image retrieval using the concept of Attribute B... more The present paper provides the information about image retrieval using the concept of Attribute Base Image Retrieval (ABIR) and (CBIR) Content Base Image Retrieval and the fusion of both method (visual and textual) which is a recent course in image retrieval researches. For CBIR we used the Scale-Invariant Feature Transform (SIFT) technique. ABIR is annotation based technique. For fusion of both ABIR and CBIR we used APRIORI algorithm.Using that algorithm we get the one result from above two results (results from ABIR and CBIR).In that algorithm it find out the relationship between visual and textual form, and then generate the result. SIFT joins two different data mining techniques to get semantically related images these are: association and clustering rules mining algorithm which is nothing but the Bag of Features (BoF). BoF methods are based on order less collection of quantized local image descriptors they remove spatial information and are therefore computationally and concept...

Background subtraction is one of the key techniques for automatic video analysis, especially in t... more Background subtraction is one of the key techniques for automatic video analysis, especially in the domain of video surveillance. Although its importance, evaluations of recent background subtraction methods with respect to the challenges of video surveillance suffer from various short comings. To address this issue, we identify the main challenges of background subtraction in the field of video surveillance. We then compare the performance of two background subtraction methods. First we subtract the background by median value method and then by histogram method. Then we see the result of both methods on Casia NLPR gait database. Keywords— Foreground detection, Median Value, Histogram, Background subtraction

In this paper, a genetic algorithm based approach for task scheduling in distributed system consi... more In this paper, a genetic algorithm based approach for task scheduling in distributed system considering dynamic load balancing is discussed. The underlying distributed system has central scheduler and task scheduling is done by this central node. Scheduling of tasks in distributed system involves deciding not only when to execute a process, but also where to execute it. A proper task scheduling will enhance the processor utilization, reduces execution time and increases system throughput. A Genetic algorithm will give the optimal solution for scheduling of task. The task scheduling is centralized and genetic algorithm is applied to central node. This task scheduling policy considers load balancing to prevent the node connected in the system from getting overloaded or become idle ever(if possible).

This paper aims to provide reliable indications of driver drowsiness describe of detecting early ... more This paper aims to provide reliable indications of driver drowsiness describe of detecting early signs of fatigue in drivers and provide method for more security and attention for driver safety problem and to investigate driver mental state related to driver safety.As soon as the driver is falling in symptons of fatigue immediate message will be given to driver.In addition of the advance technology of Surff feature extraction algorithm is also added in the system for correct detection of status of driver.The Fatigue is detected in the system by the image processing method of comparing the images(frames) in the video and by using the human features we are able to estimate the indirect way of detecting fatigue.The technique also focuses on modes of person when driving vehicle i.e awake, drowsy state or sleepy and sleep state.The system is very efficient to detect the fatigue and control

Due to the EMR (Electronic Medical Record) system there will be a rapid growth in health data col... more Due to the EMR (Electronic Medical Record) system there will be a rapid growth in health data collection. As we have already discuss in previous review paper the different work of the health care data record for maintaining the privacy and security of health care most private data. Now in this paper we are going to implement sheltered and secretive data management structure that addresses both the sheltered and secretive issues in the managementor organization of medical datainoutsourceddatabases. Theproposed framework will assure the security of data by using semantically secure encryption schemes to keep data encrypted in outsourced databases. The framework also provides a differentially-private query or uncertainty interface that can support a number of SQL queries and complicated data mining responsibilities. We are using a multiparty algorithm for this purpose. So that all the purpose is to make a secure and private management system for medical data or record storage and acces...

Data clustering is a popular approach for automatically finding classes, concepts, or groups of p... more Data clustering is a popular approach for automatically finding classes, concepts, or groups of patterns. Clustering aims at representing large datasets by a fewer number of prototypes or clusters. It brings simplicity in modeling data and thus plays a central role in the process of knowledge discovery and data mining. Data mining tasks require fast and accurate partitioning of huge datasets, which may come with a variety of attributes or features. This imposes severe computational requirements on the relevant clustering techniques. A family of bio-inspired algorithms, well-known as Swarm Intelligence (SI) has recently emerged that meets these requirements and has successfully been applied to a number of real world clustering problems. This paper looks into the use of Particle Swarm Optimization for cluster analysis. The effectiveness of Fuzzy C-means clustering provides enhanced performance and maintains more diversity in the swarm and also allows the particles to be robust to trac...

To print continuous tone images, Electrophotographic (EP) printer uses halftoning technique. Imag... more To print continuous tone images, Electrophotographic (EP) printer uses halftoning technique. Image halftoning prints the image as matrix of dots which are halftone patterns those are hardly perceived by human eyes. So scanned images obtained from such hard printed copies are normally affected by screen like artifacts and moire patterns. Scanned color Image improvement using Image descreening technique will be used to descreen color scanned images so as to remove screening patterns and Moire effects. Technique applied on grayscale images produces clean smooth regions and sharp edges from scanned halftone images. The main purpose of this paper is to study and review different methods used for scanned halftone image descreening for color and grayscale images.

In last decade, the problem of multi-label classification gain huge importance. The Multi-label c... more In last decade, the problem of multi-label classification gain huge importance. The Multi-label classification problem includes example which belongs to multiple labels. Current research on multi-label classification concentrated on supervised learning whose assumption is that huge amount of labeled training data is available. Unfortunately, in many applications labeling the training example is expensive and time consuming, mainly when it consist of more than one label. However, there are often large amount of unlabeled data is available. To solve the problem of labeling to unlabeled data transductive multi-label learning method will be used. In this paper, TRAM is used to assign multiple labels to each instance and then the result of transductive multi-label learning will be verified by Cost-sensitive Multi-label learning which will use to identify & minimize the misclassification cost of labels.

In this paper, a genetic algorithm based approach for job scheduling in distributed system consid... more In this paper, a genetic algorithm based approach for job scheduling in distributed system considering dynamic load balancing is discussed. The underlying distributed system has hierarchical structure and job scheduling is done in two levels: group level and node level. Scheduling in distributed system involves deciding not only when to execute a process, but also where to execute it. A proper job scheduling will enhance the processor utilization, reduces execution time and increases system throughput. A power of Genetic algorithm will give the optimal solution for scheduling of job. The job scheduling is centralized at each node in a hierarchy and genetic algorithm is applied to each central node. This centralized job scheduling policy considers load balancing to prevent the node connected in the system from getting overloaded or become idle ever(if possible).

This paper addresses the issue of maximizing the efficiency and scalability of RAMBased storage s... more This paper addresses the issue of maximizing the efficiency and scalability of RAMBased storage systems where in multiple objects must be retrieved per user request. The focus is on per server transaction, not per requested item. By introducing RnB, a innovative mechanism to minimize the number of servers accessed per user request, it increases the scalability and efficiency of RAMBased storage systems. In this paper, We present “Replicate and Bundle” (RnB), a method for reducing the number of transactions required to process an end user request. This method enables increasing the maximum system throughput without adding CPUs. RnB entails 1) data replication and 2) bundling of items requested from the same server into a single transaction. We use a pseudo-random object-to-server mapping for each object’s different replicas, placing the replicas on different servers for each object. During data fetch, we choose which replica to access in order to reduce the number of servers that nee...

In information technology there many types of graphs such as biological graphs ,social network gr... more In information technology there many types of graphs such as biological graphs ,social network graphs, Semantic Web graphs are presented in real world and, single vertex of any graphs always contain huge information, which can be created by a set of elements. In this paper, we will study a subgraph mining in a large graph over a large database, which creates subgraphwith help of some keywords for example name of domain and Subject. By putting all required keywords the main large graph will be created. That main graph contains rich information of different scientific searched papers, and those papers are connected in the form of graph according to year wise from which is ancient paper to new latest paper form, this connected papers show the genealogy of different seminal papers. By using the proposed method it is easy to find different seminal papers in sub graph from the main large graph. The genealogy algorithm is useful to find out quick result of searched topic, in which all foun...

In computing, there are multiple representations of same data. Finding out those duplicated data ... more In computing, there are multiple representations of same data. Finding out those duplicated data is a difficult task. Duplicate detection is the process of identifying those redundant data to reduce storage utilization and avoid the confusion. Today, duplicate detection methods need to process larger data in shorter time period. The proposed system will be having efficient approaches for finding duplicate data. The first is based on Novel progressive duplicate detection algorithms that will significantly increase the efficiency of finding duplicate files. The second is based on secure hashing algorithm which will find duplicated content by dividing the data into chunks. The best algorithm will be analysed for finding duplicate data with higher accuracy and also in a shorter time span.

International Journal of Smart Sensor and Adhoc Network.

WSN consisting of a large number of small sensors with low power transceivers can be an effective... more WSN consisting of a large number of small sensors with low power transceivers can be an effective tool for gathering data in a variety of environments. As sensor nodes are deployed in sensing field, they can help people to monitor and aggregate data. Researchers also try to find more efficient ways of utilizing limited energy of sensor node in order to give longer life time of WSNs. Network lifetime, scalability, and load balancing are important requirements for many data gathering sensor network applications. Therefore, many protocols are introduced for better performance. The lack of infrastructure and dynamic nature of mobile ad hoc networks demand new networking strategies to be implemented in order to provide efficient end-to-end communication. Some researches have been proposed to organize the network into groups called clusters and use different routing protocols for inter and intra clusters to propagate an information. But with these solutions, the network first needs to be ...

International Journal of Science and Research (IJSR)

Clustering is a crucial task in data mining. There are so many techniques for data mining which a... more Clustering is a crucial task in data mining. There are so many techniques for data mining which are continuously emerging since the concept of data mining has been taken into account. All the clustering methodologies focus on certain data or the known data. But very few methodologies focus on the clustering of uncertain data. By considering the uncertain data for clustering, the results of clustering algorithm get affected and can show any unwanted results. For example, in the sensor based applications the output can be uncertain if there is any errors in input (e.g. noise occurred during capturing information). In such situations, the algorithm must be strong enough to consider the uncertain data because the uncertain data is also important. As there is continuous increase in data accumulation in the databases we need to be more aware about to handle uncertain data. In this paper we will discuss about the data uncertainty and the various approaches to handle and cluster the uncertain data.

2016 World Conference on Futuristic Trends in Research and Innovation for Social Welfare (Startup Conclave), 2016

This paper gives the information about object matching and tracking has found important and wide ... more This paper gives the information about object matching and tracking has found important and wide applications. An object matching and tracking algorithm based on SURF (Speeded-Up Robust Feature) method is presented in this project. In existing system the SIFT (Scale-Invariant Feature Transform) method is used which have some problems like time complexity, feature extraction, occlusion problem. To overcome these problems SURF method will be used in this project. Firstly, feature points are extracted respectively from base image using SURF method. Then, a coarse-to-fine matching method is used to realize the match of SURF feature points. It shows that, compared with the frequently-used normal cross correlation method, the presented algorithm can process more complicated geometric deformations existed between images and gives high matching accuracy as compare to the matching algorithm based on SIFT feature, in the presented algorithm the processing speed is faster than other algorithm and also lower computational burden, which can meet the real-time requirements for object matching and tracking.

This paper provides the information about image retrieval using the concept of Attribute Base Ima... more This paper provides the information about image retrieval using the concept of Attribute Base Image Retrieval (ABIR) and Content Base Image Retrieval (CBIR). The retrieval method nominate in present paper utilizes the fusion of the images multimodal information (visual and textual) which is a recent trend in image retrieval researches. Image retrieval is the science of locating images from a large database or image sequences that fulfill the specified image need. An image retrieval technique is based on the Scale-Invariant Feature Transform (SIFT) method is present in this project. It joins two different data mining techniques to get semantically related images they are as: association and clustering rules mining algorithm which is nothing but the Bag of Features (BoF). BoF methods are based on order less collection of quantized local image descriptors they remove spatial information and are therefore computationally and conceptually simpler than many alternative methods.

This paper provides the information about an effective method for MRI brain image enhancement. Th... more This paper provides the information about an effective method for MRI brain image enhancement. The automatic segmentation of brain-tissue has led to the variation in the images due to different scanning and the imaging protocols which makes the image unclear and thus application is hampered. The transfer learning with weighted SVM enables training data to minimize classification errors as the classification scheme needs only a small amount of representative data. Therefore a new optimally standardized method is presented for scanned image segmentation using Transfer Learning with Weighted Support Vector Machine and then further improving the training data quality by Vector Sparse Representation using Iterative Algorithm for Quaternion Matrix Analysis over Reflexive Matrices.

2008 International Conference on Computer Science and Information Technology, 2008

... Segmentation of MRI images using intensity values is severely limited owing to in-homogeneiti... more ... Segmentation of MRI images using intensity values is severely limited owing to in-homogeneities and partial volume effects. Edge based segmentation methods suffer from unsharp edges and gaps in boundaries. A number of other methods are also used. ...

Ad hoc wireless networks are defined as the category of wireless networks that utilizes multi-hop... more Ad hoc wireless networks are defined as the category of wireless networks that utilizes multi-hop radio relaying and are capable of operating without the support of any fixed infrastructure hence, they are called infrastructure less networks. The lack of any central coordination makes them more vulnerable to attacks than wired networks. Due to some unique characteristics of MANETs, prevention methods alone are not sufficient to make them secure therefore, detection should be added as another defense before an attacker can breach the system. Network intrusion detection is the process of monitoring the events occurring in the network and analyzing them for signs of intrusions, defined as attempts to compromise the confidentiality. In this paper, we define and discuss various techniques of Intrusion Detection. We also present a description of routing protocols and types of security attacks possible in the network.

Duplicate detection is the process of identifying multiple representations of same real world ent... more Duplicate detection is the process of identifying multiple representations of same real world entities. The proposed System will compare Duplication Detection Method for best results. These methods are used for removing duplicate data and to cut redundancy. The first is based on Novel progressive duplicate detection algorithms that will significantly increase the efficiency of finding duplicates if the execution time is limited. This method maximizes the gain of overall process within the time available by reporting most results earlier than traditional approaches. The second is based on Secure Hashing Algorithm which will detect duplicate data for performing data de-duplication task to overcome the issues of time and to cut hash collision. This architecture will be useful for storage server where a huge amount of data is stored every day and software industries always looks for new developments so that they can keep their storage systems up to date and free for efficient use of it.

The present paper provides the information about image retrieval using the concept of Attribute B... more The present paper provides the information about image retrieval using the concept of Attribute Base Image Retrieval (ABIR) and (CBIR) Content Base Image Retrieval and the fusion of both method (visual and textual) which is a recent course in image retrieval researches. For CBIR we used the Scale-Invariant Feature Transform (SIFT) technique. ABIR is annotation based technique. For fusion of both ABIR and CBIR we used APRIORI algorithm.Using that algorithm we get the one result from above two results (results from ABIR and CBIR).In that algorithm it find out the relationship between visual and textual form, and then generate the result. SIFT joins two different data mining techniques to get semantically related images these are: association and clustering rules mining algorithm which is nothing but the Bag of Features (BoF). BoF methods are based on order less collection of quantized local image descriptors they remove spatial information and are therefore computationally and concept...

Background subtraction is one of the key techniques for automatic video analysis, especially in t... more Background subtraction is one of the key techniques for automatic video analysis, especially in the domain of video surveillance. Although its importance, evaluations of recent background subtraction methods with respect to the challenges of video surveillance suffer from various short comings. To address this issue, we identify the main challenges of background subtraction in the field of video surveillance. We then compare the performance of two background subtraction methods. First we subtract the background by median value method and then by histogram method. Then we see the result of both methods on Casia NLPR gait database. Keywords— Foreground detection, Median Value, Histogram, Background subtraction

In this paper, a genetic algorithm based approach for task scheduling in distributed system consi... more In this paper, a genetic algorithm based approach for task scheduling in distributed system considering dynamic load balancing is discussed. The underlying distributed system has central scheduler and task scheduling is done by this central node. Scheduling of tasks in distributed system involves deciding not only when to execute a process, but also where to execute it. A proper task scheduling will enhance the processor utilization, reduces execution time and increases system throughput. A Genetic algorithm will give the optimal solution for scheduling of task. The task scheduling is centralized and genetic algorithm is applied to central node. This task scheduling policy considers load balancing to prevent the node connected in the system from getting overloaded or become idle ever(if possible).

This paper aims to provide reliable indications of driver drowsiness describe of detecting early ... more This paper aims to provide reliable indications of driver drowsiness describe of detecting early signs of fatigue in drivers and provide method for more security and attention for driver safety problem and to investigate driver mental state related to driver safety.As soon as the driver is falling in symptons of fatigue immediate message will be given to driver.In addition of the advance technology of Surff feature extraction algorithm is also added in the system for correct detection of status of driver.The Fatigue is detected in the system by the image processing method of comparing the images(frames) in the video and by using the human features we are able to estimate the indirect way of detecting fatigue.The technique also focuses on modes of person when driving vehicle i.e awake, drowsy state or sleepy and sleep state.The system is very efficient to detect the fatigue and control

Due to the EMR (Electronic Medical Record) system there will be a rapid growth in health data col... more Due to the EMR (Electronic Medical Record) system there will be a rapid growth in health data collection. As we have already discuss in previous review paper the different work of the health care data record for maintaining the privacy and security of health care most private data. Now in this paper we are going to implement sheltered and secretive data management structure that addresses both the sheltered and secretive issues in the managementor organization of medical datainoutsourceddatabases. Theproposed framework will assure the security of data by using semantically secure encryption schemes to keep data encrypted in outsourced databases. The framework also provides a differentially-private query or uncertainty interface that can support a number of SQL queries and complicated data mining responsibilities. We are using a multiparty algorithm for this purpose. So that all the purpose is to make a secure and private management system for medical data or record storage and acces...

Data clustering is a popular approach for automatically finding classes, concepts, or groups of p... more Data clustering is a popular approach for automatically finding classes, concepts, or groups of patterns. Clustering aims at representing large datasets by a fewer number of prototypes or clusters. It brings simplicity in modeling data and thus plays a central role in the process of knowledge discovery and data mining. Data mining tasks require fast and accurate partitioning of huge datasets, which may come with a variety of attributes or features. This imposes severe computational requirements on the relevant clustering techniques. A family of bio-inspired algorithms, well-known as Swarm Intelligence (SI) has recently emerged that meets these requirements and has successfully been applied to a number of real world clustering problems. This paper looks into the use of Particle Swarm Optimization for cluster analysis. The effectiveness of Fuzzy C-means clustering provides enhanced performance and maintains more diversity in the swarm and also allows the particles to be robust to trac...

To print continuous tone images, Electrophotographic (EP) printer uses halftoning technique. Imag... more To print continuous tone images, Electrophotographic (EP) printer uses halftoning technique. Image halftoning prints the image as matrix of dots which are halftone patterns those are hardly perceived by human eyes. So scanned images obtained from such hard printed copies are normally affected by screen like artifacts and moire patterns. Scanned color Image improvement using Image descreening technique will be used to descreen color scanned images so as to remove screening patterns and Moire effects. Technique applied on grayscale images produces clean smooth regions and sharp edges from scanned halftone images. The main purpose of this paper is to study and review different methods used for scanned halftone image descreening for color and grayscale images.

In last decade, the problem of multi-label classification gain huge importance. The Multi-label c... more In last decade, the problem of multi-label classification gain huge importance. The Multi-label classification problem includes example which belongs to multiple labels. Current research on multi-label classification concentrated on supervised learning whose assumption is that huge amount of labeled training data is available. Unfortunately, in many applications labeling the training example is expensive and time consuming, mainly when it consist of more than one label. However, there are often large amount of unlabeled data is available. To solve the problem of labeling to unlabeled data transductive multi-label learning method will be used. In this paper, TRAM is used to assign multiple labels to each instance and then the result of transductive multi-label learning will be verified by Cost-sensitive Multi-label learning which will use to identify & minimize the misclassification cost of labels.

In this paper, a genetic algorithm based approach for job scheduling in distributed system consid... more In this paper, a genetic algorithm based approach for job scheduling in distributed system considering dynamic load balancing is discussed. The underlying distributed system has hierarchical structure and job scheduling is done in two levels: group level and node level. Scheduling in distributed system involves deciding not only when to execute a process, but also where to execute it. A proper job scheduling will enhance the processor utilization, reduces execution time and increases system throughput. A power of Genetic algorithm will give the optimal solution for scheduling of job. The job scheduling is centralized at each node in a hierarchy and genetic algorithm is applied to each central node. This centralized job scheduling policy considers load balancing to prevent the node connected in the system from getting overloaded or become idle ever(if possible).

This paper addresses the issue of maximizing the efficiency and scalability of RAMBased storage s... more This paper addresses the issue of maximizing the efficiency and scalability of RAMBased storage systems where in multiple objects must be retrieved per user request. The focus is on per server transaction, not per requested item. By introducing RnB, a innovative mechanism to minimize the number of servers accessed per user request, it increases the scalability and efficiency of RAMBased storage systems. In this paper, We present “Replicate and Bundle” (RnB), a method for reducing the number of transactions required to process an end user request. This method enables increasing the maximum system throughput without adding CPUs. RnB entails 1) data replication and 2) bundling of items requested from the same server into a single transaction. We use a pseudo-random object-to-server mapping for each object’s different replicas, placing the replicas on different servers for each object. During data fetch, we choose which replica to access in order to reduce the number of servers that nee...

In information technology there many types of graphs such as biological graphs ,social network gr... more In information technology there many types of graphs such as biological graphs ,social network graphs, Semantic Web graphs are presented in real world and, single vertex of any graphs always contain huge information, which can be created by a set of elements. In this paper, we will study a subgraph mining in a large graph over a large database, which creates subgraphwith help of some keywords for example name of domain and Subject. By putting all required keywords the main large graph will be created. That main graph contains rich information of different scientific searched papers, and those papers are connected in the form of graph according to year wise from which is ancient paper to new latest paper form, this connected papers show the genealogy of different seminal papers. By using the proposed method it is easy to find different seminal papers in sub graph from the main large graph. The genealogy algorithm is useful to find out quick result of searched topic, in which all foun...

In computing, there are multiple representations of same data. Finding out those duplicated data ... more In computing, there are multiple representations of same data. Finding out those duplicated data is a difficult task. Duplicate detection is the process of identifying those redundant data to reduce storage utilization and avoid the confusion. Today, duplicate detection methods need to process larger data in shorter time period. The proposed system will be having efficient approaches for finding duplicate data. The first is based on Novel progressive duplicate detection algorithms that will significantly increase the efficiency of finding duplicate files. The second is based on secure hashing algorithm which will find duplicated content by dividing the data into chunks. The best algorithm will be analysed for finding duplicate data with higher accuracy and also in a shorter time span.

International Journal of Smart Sensor and Adhoc Network.

WSN consisting of a large number of small sensors with low power transceivers can be an effective... more WSN consisting of a large number of small sensors with low power transceivers can be an effective tool for gathering data in a variety of environments. As sensor nodes are deployed in sensing field, they can help people to monitor and aggregate data. Researchers also try to find more efficient ways of utilizing limited energy of sensor node in order to give longer life time of WSNs. Network lifetime, scalability, and load balancing are important requirements for many data gathering sensor network applications. Therefore, many protocols are introduced for better performance. The lack of infrastructure and dynamic nature of mobile ad hoc networks demand new networking strategies to be implemented in order to provide efficient end-to-end communication. Some researches have been proposed to organize the network into groups called clusters and use different routing protocols for inter and intra clusters to propagate an information. But with these solutions, the network first needs to be ...

International Journal of Science and Research (IJSR)

Clustering is a crucial task in data mining. There are so many techniques for data mining which a... more Clustering is a crucial task in data mining. There are so many techniques for data mining which are continuously emerging since the concept of data mining has been taken into account. All the clustering methodologies focus on certain data or the known data. But very few methodologies focus on the clustering of uncertain data. By considering the uncertain data for clustering, the results of clustering algorithm get affected and can show any unwanted results. For example, in the sensor based applications the output can be uncertain if there is any errors in input (e.g. noise occurred during capturing information). In such situations, the algorithm must be strong enough to consider the uncertain data because the uncertain data is also important. As there is continuous increase in data accumulation in the databases we need to be more aware about to handle uncertain data. In this paper we will discuss about the data uncertainty and the various approaches to handle and cluster the uncertain data.

2016 World Conference on Futuristic Trends in Research and Innovation for Social Welfare (Startup Conclave), 2016

This paper gives the information about object matching and tracking has found important and wide ... more This paper gives the information about object matching and tracking has found important and wide applications. An object matching and tracking algorithm based on SURF (Speeded-Up Robust Feature) method is presented in this project. In existing system the SIFT (Scale-Invariant Feature Transform) method is used which have some problems like time complexity, feature extraction, occlusion problem. To overcome these problems SURF method will be used in this project. Firstly, feature points are extracted respectively from base image using SURF method. Then, a coarse-to-fine matching method is used to realize the match of SURF feature points. It shows that, compared with the frequently-used normal cross correlation method, the presented algorithm can process more complicated geometric deformations existed between images and gives high matching accuracy as compare to the matching algorithm based on SIFT feature, in the presented algorithm the processing speed is faster than other algorithm and also lower computational burden, which can meet the real-time requirements for object matching and tracking.