Sonia Vatta | Rayat Bahra University (original) (raw)
Papers by Sonia Vatta
Wireless sensor network is a network that consists of tiny, complex and large number of sensors a... more Wireless sensor network is a network that consists of tiny, complex and large number of sensors and at least one base station or sink node. Most challenging issue in wireless sensor network is the limited battery power of sensor nodes used in the network. To increase the energy of sensor nodes, energy is preferably dispensed throughout the wireless sensor network. So the key to enhance the life time of the network is to design effective and energy aware protocols. Routing protocol can be network structure based or protocol operation based. In this paper, E-DEC (Enhanced Deterministic Energy Efficient) Protocol is proposed. It assumes WSN containing four energy levels of nodes. Here, Cluster Heads (CHs) are elected on the bases of residual or remaining energy of nodes. Experimental result shows a better performance with respect to energy consumption, which is reflected in the better network lifetime in both homogeneous and heterogeneous settings when compared with the existing protoc...
In this paper, description of a compression algorithm based o n steganography has been narrated. ... more In this paper, description of a compression algorithm based o n steganography has been narrated. The compression algorithm has been used to develop an application which will help the users to hide large size text documents inside small size images. Maximum bits to be hidden per pixel can be increased to eight with the help of the developed compression application. After hiding the data inside an image, there appears to be no visible distortion at all. Also the application is compatible with all the documents and image formats. The developed application automatically converts the output stego image in bmp format.
— Green computing is the study and practice of efficient and eco-friendly computing. Green comput... more — Green computing is the study and practice of efficient and eco-friendly computing. Green computing is also called as green technology. The principle behind energy efficient coding is to save power by getting software to make less use of the hardware, rather than continuing to run the same code on hardware that uses less power. Green computing is the environmentally responsible use of computers and related resources. Such practices include the implementation of energy-efficient central processing units (CPUs), servers and peripherals as well as reduced resource consumption and proper disposal of electronic waste (e-waste).Green computing is also necessary for the future generation also. This work includes the use of green computing in today’s world and how the environment problems can reduced using green computing and how to protect the future by using the green technology.
Abstract- As we know the computers are widely used in every field either it is of geography, medi... more Abstract- As we know the computers are widely used in every field either it is of geography, medical, pharmacy, astrology. Astronomy and so on. The ongoing advancements in all these fields require a big database and the place from where this data is retrieved easily to use. The data is some time is in hierarchal format. But the array of memory that we use to save information is in only in 2-D, we have trees for such information in data structure. Q- Tree or Quad tree one of the ways of representing data in the memory. The problem is of representing data in this tree so that one can do searching, insertion and deletion in a fastest manner. With the increase in traversing and searching the performance of the computer too increases
Face recognition is a system that identifies human faces through complex computational techniques... more Face recognition is a system that identifies human faces through complex computational techniques. The paper explains two different algorithms for feature extraction. These are Principal Component Analysis and Fisher Faces algorithm. It then explains how images can be recognized using a backpropagation algorithm on a feed forward neural network. Two training databases one containing 20 images and the other containing 80 images are used to test the proposed techniques. Later the results are compared and tabulated.
In recent years, automotive manufacturers have equipped their vehicles with innovative Advanced D... more In recent years, automotive manufacturers have equipped their vehicles with innovative Advanced Driver Assistance Systems (ADAS) to ease driving and avoid dangerous situations, such as unintended lane departures or collisions with other road users, like vehicles and pedestrians. To this end, ADAS at the cutting edge are equipped with cameras to sense the vehicle surrounding. This dissertation investigates the techniques for monocular vision based vehicle detection. A system that can robustly detect and track vehicles in images. The system consists of three major modules: Histogram of oriented gradient (HOG) is used as the main feature descriptor which is shape oriented, a machine learning part based on support vector machine (SVM) for vehicle verification, to make the system biological human eye driven lastly a technique is applied for texture analysis by applying the concept of gray level co-occurrence matrix (GLCM).More specifically, we are interested in detection of cars from dif...
In this paper we discuss the issue of privacy preserving data mining and present the technique th... more In this paper we discuss the issue of privacy preserving data mining and present the technique that provide the privacy on data mining application. We provide an overview of the different techniques and how they relate to one another. We used the asymmetric encryption to provide the privacy and used the RSA encryption to encrypt the data. We also present a client server architecture that connects to the multiple clients. Server need to receive data from client .Connect server to a data base and enter the data received from the client into database with client id. Our proposed protocol is to encrypt the data so we used the encryption technique to encrypt the data, we also used homomorphic encryption to secure the information. Without security our data may stand compromised.
Learning that is based on induction is the inductive learning. Decision tree algorithms are very ... more Learning that is based on induction is the inductive learning. Decision tree algorithms are very famous in inductive learning. These kinds of algorithm use inductive methods for the appropriate classification of the objects with the given attributes. These algorithms are very beneficial in the classification of the objects and are mainly used in expert systems. In this paper the ID3 decision tree learning algorithm is used to find out whether there are any changes in the present decision rules for issuing of PUCC (Pollution under Control Certificate) when some new attributes are added. Here the three studies are done regarding the issuing of PUCC and in each study a new attribute is added to the dataset to get the decision rules as resultant. The algorithm is implemented in the java language.
Cloud computing, the long-held dream of computing as a utility, has finally came to realization. ... more Cloud computing, the long-held dream of computing as a utility, has finally came to realization. Cloud computing has the potential to transform a large part of the IT industry, making software even more attractive as a service and shaping the way IT hardware is designed and purchased. Developers with innovative ideas for new Internet services no longer require the large capital outlays in hardware to deploy their service or the human expense to operate it. These cloud applications use large data centres and powerful servers that host Web applications and Web services.This report gives an overview of what cloud computing means, its history along with the advantages and disadvantages. Finally the different types of Cloud computing services commonly referred to as Software as a Service (SaaS), Platform as a Service (PaaS) and Infrastructure as a Service (IaaS) are explained along with some examples to illustrate how they all work. This report also provides some guidance on situations w...
Near-infrared (NIR) is a part of the electromagnetic spectrum with an approximate wavelength rang... more Near-infrared (NIR) is a part of the electromagnetic spectrum with an approximate wavelength range of 700 nm to 1100 nm. Mostly all the digital cameras can capture NIR spectrum by doing a slight modifications. To take advantage of both NIR and visible information, it is necessary to simultaneously capture NIR and color images for each scene. RGB and NIR cues have been successfully combined in many applications like dehazing, dark flash photography, and scene categorization. NIR can be widely used in semantic segmentation and many other image processing techniques.
Face recognition is a system that identifies human faces from an image database or from a video f... more Face recognition is a system that identifies human faces from an image database or from a video frame . The paper presents a literature review on face recognition approaches. It then explains two different algorithms for feature extraction which are Principal Component Analysis and Fisher Faces algorithm. It also explains how images can be recognized using a Backpropagation algorithm on a Feedforward neural network and Minimum Euclidean Distance. Keywords—eigenfaces, fisherfaces, neural network, backpropagation, minimum euclidean distance
International journal of scientific research in science, engineering and technology, 2015
This research work is done to optimize the search in hierarchal database by using Quad Tree. Quad... more This research work is done to optimize the search in hierarchal database by using Quad Tree. Quad Tree is a tree which consists of a node having 4 sub child nodes.[1] Quad Tree is used in many fields such as computer graphics to find out a desired pixel in graphics and collision detection in 3-D gaming. There are various other applications of quad tree in different fields like weather forecasting, geographical survey and study of mutation rate with a change of environment. The main focus of this work was optimizing the search for which the main areas of concentration were various searching techniques using Quad Tree and finding out the best solution. To fulfill the purpose, a new approach is designed which results in better representation of data and enhancing the speed of search. This work is done on the basis of the value of the root node. Further based on this value, the ranges of nodes at each level are calculated. The results are also included with snapshots and comparison is g...
IOSR Journal of Engineering, 2014
A WSN is a specialized wireless network made up of large number of sensors and at least one base ... more A WSN is a specialized wireless network made up of large number of sensors and at least one base station. The foremost or the main difference between WSN and the traditional wireless networks is that sensors are extremely sensitive to energy consumption. Energy saving in the crucial issue in designing the wireless sensor networks [12]... Since the radio transmission and reception consumes more energy, one of the most or the main significant issue in wireless sensor network is the inherent limited battery power within network sensor nodes. It is preferable to dispense the energy throughout the wireless sensor network so to maximize the lifetime of sensor nodes. So it is essential to design effective and energy aware protocols in order to enhance the life time of the network. A wireless sensor network may have network structure based or protocol operation based routing protocol. In this paper, a review on hierarchical based routing protocol which is further a sub-type of the network structure based routing protocol in WSNs is carried out. Major issues which are considered in WSNs are Energy consumption and network life time.
Inductive learning is the learning that is based on induction. In inductive learningDecision tree... more Inductive learning is the learning that is based on induction. In inductive learningDecision tree algorithms are very famous. For the appropriate classification of the objects with the given attributes inductive methods use these algorithms basically. These algorithms are very important in the classification of the objects. That is why many of these algorithms are used in the intelligent systems as well. In this paper the ID3 decision tree learning algorithm is implemented with the help of an example which includes the training set of two weeks. The basic calculations are used to calculate the classification related to the training set used.The resultant of the work will be the classified decision tree and the decision rules. The algorithm is implemented in the java language.
Content-Based Image Retrieval is the field of digital image processing that has been used for the... more Content-Based Image Retrieval is the field of digital image processing that has been used for the extraction of valuable information from huge datasets. In this process of CBIR, images have been extracted from huge datasets based on content available in the images. Various types of images are available under digital imaging. Different types of features have to be computed so that images can be extracted from the datasets on the basis of features. In this paper, various approaches have been discussed that has been used for the extraction of features based on content. Color, shape and texture based features have been extracted from the digital images so that relevant information available in the datasets can be extracted. This paper comprises review about various approaches of feature extraction from digital images. On the basis of review of these approaches, one can analyze best approach for feature extraction.
With the advancement of technology, social media is a basic need of the modern world. Due to incr... more With the advancement of technology, social media is a basic need of the modern world. Due to increasing the popularity of social media, cyberbullying is also becoming a big concern. If a person harasses another person using digital appliances like mobile phones, internet or emails is called cyberbullying. The bully person can be very intelligent, he knows all the schemes to create problem to the victim and can easily hide his identity. The cyberbullying is very difficult to identify because the bully messages are very less as compared to the normal messages. The classification and feature extraction of the data set is very difficult. But Data mining is the best and popular technology for classification and regression. It is the process by which useful information and patterns are extracted from a large amount of data stored in the databases. Data Mining is also known as knowledge discovery process as knowledge is being extracted or patterns are analyzed which is very useful to collect data. It is very important to note that the exactness and availability of the results will rely on the standard of information investigation and the nature of the suspicious user. In the existing work, the Naïve Bayes classifier has been used on the cyberbullying. But there is huge no. of limitations for classifying the cyberbullying like accuracy, precision time, recall, execution time. In this review paper, the Random Forest classifier will be described for the cyberbullying. It is expected that Random Forest classifier will high accuracy and high sensitivity as compared to Naïve Bayes.
This paper presents a review of steganography and various steganography techniques used for data ... more This paper presents a review of steganography and various steganography techniques used for data compression. The purpose is to have a deep study of various steganographic techniques used for data compression. The main objective is to find out a technique, which can hide a large amount of data. To fulfil the purpose, various researches and projects done earlier are taken into consideration.
Wireless sensor network is a network that consists of tiny, complex and large number of sensors a... more Wireless sensor network is a network that consists of tiny, complex and large number of sensors and at least one base station or sink node. Most challenging issue in wireless sensor network is the limited battery power of sensor nodes used in the network. To increase the energy of sensor nodes, energy is preferably dispensed throughout the wireless sensor network. So the key to enhance the life time of the network is to design effective and energy aware protocols. Routing protocol can be network structure based or protocol operation based. In this paper, E-DEC (Enhanced Deterministic Energy Efficient) Protocol is proposed. It assumes WSN containing four energy levels of nodes. Here, Cluster Heads (CHs) are elected on the bases of residual or remaining energy of nodes. Experimental result shows a better performance with respect to energy consumption, which is reflected in the better network lifetime in both homogeneous and heterogeneous settings when compared with the existing protoc...
In this paper, description of a compression algorithm based o n steganography has been narrated. ... more In this paper, description of a compression algorithm based o n steganography has been narrated. The compression algorithm has been used to develop an application which will help the users to hide large size text documents inside small size images. Maximum bits to be hidden per pixel can be increased to eight with the help of the developed compression application. After hiding the data inside an image, there appears to be no visible distortion at all. Also the application is compatible with all the documents and image formats. The developed application automatically converts the output stego image in bmp format.
— Green computing is the study and practice of efficient and eco-friendly computing. Green comput... more — Green computing is the study and practice of efficient and eco-friendly computing. Green computing is also called as green technology. The principle behind energy efficient coding is to save power by getting software to make less use of the hardware, rather than continuing to run the same code on hardware that uses less power. Green computing is the environmentally responsible use of computers and related resources. Such practices include the implementation of energy-efficient central processing units (CPUs), servers and peripherals as well as reduced resource consumption and proper disposal of electronic waste (e-waste).Green computing is also necessary for the future generation also. This work includes the use of green computing in today’s world and how the environment problems can reduced using green computing and how to protect the future by using the green technology.
Abstract- As we know the computers are widely used in every field either it is of geography, medi... more Abstract- As we know the computers are widely used in every field either it is of geography, medical, pharmacy, astrology. Astronomy and so on. The ongoing advancements in all these fields require a big database and the place from where this data is retrieved easily to use. The data is some time is in hierarchal format. But the array of memory that we use to save information is in only in 2-D, we have trees for such information in data structure. Q- Tree or Quad tree one of the ways of representing data in the memory. The problem is of representing data in this tree so that one can do searching, insertion and deletion in a fastest manner. With the increase in traversing and searching the performance of the computer too increases
Face recognition is a system that identifies human faces through complex computational techniques... more Face recognition is a system that identifies human faces through complex computational techniques. The paper explains two different algorithms for feature extraction. These are Principal Component Analysis and Fisher Faces algorithm. It then explains how images can be recognized using a backpropagation algorithm on a feed forward neural network. Two training databases one containing 20 images and the other containing 80 images are used to test the proposed techniques. Later the results are compared and tabulated.
In recent years, automotive manufacturers have equipped their vehicles with innovative Advanced D... more In recent years, automotive manufacturers have equipped their vehicles with innovative Advanced Driver Assistance Systems (ADAS) to ease driving and avoid dangerous situations, such as unintended lane departures or collisions with other road users, like vehicles and pedestrians. To this end, ADAS at the cutting edge are equipped with cameras to sense the vehicle surrounding. This dissertation investigates the techniques for monocular vision based vehicle detection. A system that can robustly detect and track vehicles in images. The system consists of three major modules: Histogram of oriented gradient (HOG) is used as the main feature descriptor which is shape oriented, a machine learning part based on support vector machine (SVM) for vehicle verification, to make the system biological human eye driven lastly a technique is applied for texture analysis by applying the concept of gray level co-occurrence matrix (GLCM).More specifically, we are interested in detection of cars from dif...
In this paper we discuss the issue of privacy preserving data mining and present the technique th... more In this paper we discuss the issue of privacy preserving data mining and present the technique that provide the privacy on data mining application. We provide an overview of the different techniques and how they relate to one another. We used the asymmetric encryption to provide the privacy and used the RSA encryption to encrypt the data. We also present a client server architecture that connects to the multiple clients. Server need to receive data from client .Connect server to a data base and enter the data received from the client into database with client id. Our proposed protocol is to encrypt the data so we used the encryption technique to encrypt the data, we also used homomorphic encryption to secure the information. Without security our data may stand compromised.
Learning that is based on induction is the inductive learning. Decision tree algorithms are very ... more Learning that is based on induction is the inductive learning. Decision tree algorithms are very famous in inductive learning. These kinds of algorithm use inductive methods for the appropriate classification of the objects with the given attributes. These algorithms are very beneficial in the classification of the objects and are mainly used in expert systems. In this paper the ID3 decision tree learning algorithm is used to find out whether there are any changes in the present decision rules for issuing of PUCC (Pollution under Control Certificate) when some new attributes are added. Here the three studies are done regarding the issuing of PUCC and in each study a new attribute is added to the dataset to get the decision rules as resultant. The algorithm is implemented in the java language.
Cloud computing, the long-held dream of computing as a utility, has finally came to realization. ... more Cloud computing, the long-held dream of computing as a utility, has finally came to realization. Cloud computing has the potential to transform a large part of the IT industry, making software even more attractive as a service and shaping the way IT hardware is designed and purchased. Developers with innovative ideas for new Internet services no longer require the large capital outlays in hardware to deploy their service or the human expense to operate it. These cloud applications use large data centres and powerful servers that host Web applications and Web services.This report gives an overview of what cloud computing means, its history along with the advantages and disadvantages. Finally the different types of Cloud computing services commonly referred to as Software as a Service (SaaS), Platform as a Service (PaaS) and Infrastructure as a Service (IaaS) are explained along with some examples to illustrate how they all work. This report also provides some guidance on situations w...
Near-infrared (NIR) is a part of the electromagnetic spectrum with an approximate wavelength rang... more Near-infrared (NIR) is a part of the electromagnetic spectrum with an approximate wavelength range of 700 nm to 1100 nm. Mostly all the digital cameras can capture NIR spectrum by doing a slight modifications. To take advantage of both NIR and visible information, it is necessary to simultaneously capture NIR and color images for each scene. RGB and NIR cues have been successfully combined in many applications like dehazing, dark flash photography, and scene categorization. NIR can be widely used in semantic segmentation and many other image processing techniques.
Face recognition is a system that identifies human faces from an image database or from a video f... more Face recognition is a system that identifies human faces from an image database or from a video frame . The paper presents a literature review on face recognition approaches. It then explains two different algorithms for feature extraction which are Principal Component Analysis and Fisher Faces algorithm. It also explains how images can be recognized using a Backpropagation algorithm on a Feedforward neural network and Minimum Euclidean Distance. Keywords—eigenfaces, fisherfaces, neural network, backpropagation, minimum euclidean distance
International journal of scientific research in science, engineering and technology, 2015
This research work is done to optimize the search in hierarchal database by using Quad Tree. Quad... more This research work is done to optimize the search in hierarchal database by using Quad Tree. Quad Tree is a tree which consists of a node having 4 sub child nodes.[1] Quad Tree is used in many fields such as computer graphics to find out a desired pixel in graphics and collision detection in 3-D gaming. There are various other applications of quad tree in different fields like weather forecasting, geographical survey and study of mutation rate with a change of environment. The main focus of this work was optimizing the search for which the main areas of concentration were various searching techniques using Quad Tree and finding out the best solution. To fulfill the purpose, a new approach is designed which results in better representation of data and enhancing the speed of search. This work is done on the basis of the value of the root node. Further based on this value, the ranges of nodes at each level are calculated. The results are also included with snapshots and comparison is g...
IOSR Journal of Engineering, 2014
A WSN is a specialized wireless network made up of large number of sensors and at least one base ... more A WSN is a specialized wireless network made up of large number of sensors and at least one base station. The foremost or the main difference between WSN and the traditional wireless networks is that sensors are extremely sensitive to energy consumption. Energy saving in the crucial issue in designing the wireless sensor networks [12]... Since the radio transmission and reception consumes more energy, one of the most or the main significant issue in wireless sensor network is the inherent limited battery power within network sensor nodes. It is preferable to dispense the energy throughout the wireless sensor network so to maximize the lifetime of sensor nodes. So it is essential to design effective and energy aware protocols in order to enhance the life time of the network. A wireless sensor network may have network structure based or protocol operation based routing protocol. In this paper, a review on hierarchical based routing protocol which is further a sub-type of the network structure based routing protocol in WSNs is carried out. Major issues which are considered in WSNs are Energy consumption and network life time.
Inductive learning is the learning that is based on induction. In inductive learningDecision tree... more Inductive learning is the learning that is based on induction. In inductive learningDecision tree algorithms are very famous. For the appropriate classification of the objects with the given attributes inductive methods use these algorithms basically. These algorithms are very important in the classification of the objects. That is why many of these algorithms are used in the intelligent systems as well. In this paper the ID3 decision tree learning algorithm is implemented with the help of an example which includes the training set of two weeks. The basic calculations are used to calculate the classification related to the training set used.The resultant of the work will be the classified decision tree and the decision rules. The algorithm is implemented in the java language.
Content-Based Image Retrieval is the field of digital image processing that has been used for the... more Content-Based Image Retrieval is the field of digital image processing that has been used for the extraction of valuable information from huge datasets. In this process of CBIR, images have been extracted from huge datasets based on content available in the images. Various types of images are available under digital imaging. Different types of features have to be computed so that images can be extracted from the datasets on the basis of features. In this paper, various approaches have been discussed that has been used for the extraction of features based on content. Color, shape and texture based features have been extracted from the digital images so that relevant information available in the datasets can be extracted. This paper comprises review about various approaches of feature extraction from digital images. On the basis of review of these approaches, one can analyze best approach for feature extraction.
With the advancement of technology, social media is a basic need of the modern world. Due to incr... more With the advancement of technology, social media is a basic need of the modern world. Due to increasing the popularity of social media, cyberbullying is also becoming a big concern. If a person harasses another person using digital appliances like mobile phones, internet or emails is called cyberbullying. The bully person can be very intelligent, he knows all the schemes to create problem to the victim and can easily hide his identity. The cyberbullying is very difficult to identify because the bully messages are very less as compared to the normal messages. The classification and feature extraction of the data set is very difficult. But Data mining is the best and popular technology for classification and regression. It is the process by which useful information and patterns are extracted from a large amount of data stored in the databases. Data Mining is also known as knowledge discovery process as knowledge is being extracted or patterns are analyzed which is very useful to collect data. It is very important to note that the exactness and availability of the results will rely on the standard of information investigation and the nature of the suspicious user. In the existing work, the Naïve Bayes classifier has been used on the cyberbullying. But there is huge no. of limitations for classifying the cyberbullying like accuracy, precision time, recall, execution time. In this review paper, the Random Forest classifier will be described for the cyberbullying. It is expected that Random Forest classifier will high accuracy and high sensitivity as compared to Naïve Bayes.
This paper presents a review of steganography and various steganography techniques used for data ... more This paper presents a review of steganography and various steganography techniques used for data compression. The purpose is to have a deep study of various steganographic techniques used for data compression. The main objective is to find out a technique, which can hide a large amount of data. To fulfil the purpose, various researches and projects done earlier are taken into consideration.