Vellore Institute of Technology (VIT) is a private university located in&nbsp;Tamil Nadu, India. Founded in 1984, as Vellore Engineering College, the institution offers 20 undergraduate, 34 postgraduate, four integrated and four research programs. It has campuses in Vellore, Amravati, Bhopal and Chennai.

VIT is one of the top ranked private universities in India according to NIRF, THE and QS Rankings.&nbsp;Govt. of India has recognized&nbsp;VIT, Vellore as an&nbsp;Institution of Eminence. This has allowed VIT to take independent quality initiatives and move up in world ranking.

&nbsp;

&nbsp;

VIT University

Image retrieval is broadly classified into two parts—Description-based image retrieval and content-based image retrieval. DBIR involves annotating the images and then using text retrieval techniques to retrieve them. This requires vast amount of labor and is expensive for large image databases. Semantic gap is another major issue in DBIR. Different people may perceive the contents of the image differently and hence annotate it differently, e.g., a bright shiny silver car for one person can be boring dull gray car for another. Thus, no objective keywords for search can be defined, whereas CBIR aims at indexing and retrieving images based on their visual contents like color, texture, and shape features. For a content-based image retrieval system, the performance is dependent on how the feature vector is represented and what similarity metrics is chosen. The disadvantage with visual content descriptor is that it is very domain specific, e.g., color correlogram and color histogram for color feature extraction, Tamura and GLCM for texture feature extraction, local binary pattern and HOG for face recognition, SIFT and SURF for object detection, etc. Hence, convolution neural network which is not domain specific would be a better approach. © 2021, Springer Nature Singapore Pte Ltd.

World University root node

Computer Science

School of Computer Science and Engineering

Chennai Campus

Lecture Notes in Electrical Engineering

Image-Based Recommendation Engine Using VGG Model

Machine learning facilitates predictive maintenance due to the advantages it holds over traditional methods of maintaining semi-conductor devices such as preventive and breakdown maintenance. Several predictive models using machine learning on the Semiconductor Manufacturing process dataset (SECOM) will be applied in this paper. The dataset contains the information related to semiconductor manufacturing process, with the attributes corresponding to signals collected from semiconductor devices. Due to the high-dimensionality of the data and class imbalance problem in the SECOM dataset, it poses several challenges related to data pre-processing, which is an essential step incorporated in this work while applying various machine learning models. Comparison and analysis of various predictive machine learning classification models were carried out based on the performance metrics like, accuracy and Receiver Operating Characteristic (ROC) curve. © Published under licence by IOP Publishing Ltd.

Fulltext

IOP Conference Series: Materials Science and Engineering

Predictive Maintenance using Machine Learning Based Classification Models

Introduction: Monitoring neurophysiological signals and neuroimaging helps the researchers to identify the relations between the major disorders in neurodevelopmental and behavioural changes. Specifically, the brain typical pattern that leads to Autism Spectrum Disorders (ASD) can be analyzed in children with the signals effectively. Objectives: Early identification of Autism Spectrum Disorders syndrome helps the parents to decide about the therapies and treatments required for improving their day to day activities. Materials and Methods: Different analyses performed by various researchers using machine learning algorithms using the EEG and ECG signals have been presented. Different techniques and devices used for acquiring the data through various experiments are also presented. Results: This paper presented a few techniques followed by various researchers; accuracy, sensitivity, recall achieved by the methodologies have been presented. Conclusion: Data collection plays a major role in analyzing the Autism Spectrum Disorders where most of the researches find it very difficult and early diagnosis also plays a major role in the improvement the Autism Spectrum Disorders children. © IJCRR.

International Journal of Current Research and Review

Autistic disorder analysis among children and adult

At present the low cost, low power and collaborative feature of Wireless Sensor Network (WSN) is becoming a popular communication technology in smart grid including power generation, transmission and distribution. Among these, the health monitoring of wind power generation system has emerged as one of the many possible applications of WSNs. However the harsh environmental condition of wind farm application brings node deployment as a major design issue in WSN which is well associated with coverage and connectivity issues. Hence the research objective here is twofold. Firstly the sensor nodes are placed optimally in the key components of the wind turbines to improve target coverage. The adjacent turbine span varies with several hundred meters apart which results in independent wireless sensor sub-networks. Connectivity among these sub-networks is a second vital issue, which is guaranteed by joining all the independent sub-networks with the base station by placing minimum number of relay nodes. Hence the connectivity problem is considered as Relay Node Deployment Problem. Connectivity is obtained in this work by bio-inspired Ant Colony Optimization (ACO) algorithm. ACO is further enhanced as ACO-Intelligent Movement, by introducing intelligent movement mechanism. The goal of this approach is to optimize number of relay nodes, decrease deployment cost and to bring up network connectivity. The performance of our novel deployment approach is validated through extensive simulation results. © 2019, Springer Science+Business Media, LLC, part of Springer Nature.

Wireless Personal Communications

Optimized Node Deployment in Wireless Sensor Network for Smart Grid Application

In recent times, health problems like cardiac failure, lung failures and heart related diseases are arising day by day at a very high rate. To address these problems, a sustained health monitoring becomes imperative. By regular health monitoring, diseases can be diagnosed and treated early. A new concept which is emerging is by monitoring the health of a patient wirelessly. In this paper, we report the use of a body temperature sensor and a heartbeat sensor for an IoT-based health monitoring system. The set up involved relaying of data from the sensors to an Arduino which was then sent to the cloud via an Ethernet module. The collected data were further processed by applying various algorithms and classification was done. We present the results obtained from the processing of the data using various algorithms-Naive Bayes, Random Forest Tree and SVM-and their comparison based on accuracy, precision and classification time. © BEIESP.

International Journal of Recent Technology and Engineering

Data analytics on IoT-based health monitoring system

In today's world, warfare is an important factor in any nation's security. One of the important and vital roles is played by the army soldiers. There are many concerns regarding the safety of soldiers. So for their security purpose, many instruments are mounted on them to view their health status as well as their real time location. Bio-sensor systems comprise various types of small physiological sensors, transmission modules and processing capabilities, and can thus facilitate low-cost wearable unobtrusive solutions for health monitoring. This paper gives an ability to track the location and monitor health of the soldiers in real time who become lost and get injured in the battlefield. It helps to minimize the time, search and rescue operation efforts of army control unit. This system enables to army control unit to track the location and monitor health of soldiers using GPS module and wireless body area sensor networks (WBASNs), such as temperature sensor, heart beat sensor, etc. The data coming from sensors and GPS receiver will be transmitted wirelessly using ZigBee module among the fellow soldiers. Furthermore, LoRaWAN network infrastructure has been proposed to be used between the squadron leader and the control unit in high altitude warzones where cellular network coverage is either absent or does not allow data transmission. The collected data will be uploaded on the cloud for further data analysis and predictions using K-Means Clustering algorithm. © 2018 The Authors. Published by Elsevier Ltd.

VIT Unitversity

Procedia Computer Science

IoT-based Healthcare Monitoring System for War Soldiers using Machine Learning

Wireless sensor network (WSN) is recognized a potential monitoring, controlling and communication tool for future power grid, the smart grid. Particularly for wind-based power generation application, the multifunctional wireless sensor nodes are positioned at the possible locations of the wind mill for monitoring. Each wind turbine is said to be a sub-network, where the sensor nodes are deployed at the potential data points to accomplish coverage. Connectivity is then a critical issue because the adjacent wind turbines are very far from each other. Hence, the sensor nodes between adjacent turbines will not be within the communication range. Relay nodes are employed between the turbines to bring up connectivity. We propose a Deterministic Relay node deployment algorithm based on Fermat point and Convex hull (DRFC) to find optimal number of relay nodes. DRFC finds the convex hull and Fermat points recursively until it reaches the centre of target region. The relay nodes are deployed in the Fermat points and also between the Fermat points and the segments to establish connectivity. The DRFC algorithm is analysed mathematically with respect to optimum number of relay node and connectivity. © Springer Nature Singapore Pte Ltd. 2018.

Lecture Notes in Electrical Engineering Advances in Smart Grid and Renewable Energy

Deterministic Node Deployment for Connected Target Coverage Problem in Heterogeneous Wireless Sensor Networks for Monitoring Wind Farm

An Information and Communication Technology (ICT) is an emerging field in the healthcare system applications for efficient health monitoring, disease detection and treatment process. Current venture of Internet of Things (IoT) in ICT has paved the way for continuous heath monitoring remotely and wirelessly. By regular health monitoring, diseases can be diagnosed and treated very early. In this re- search work, a remote healthcare monitoring system is proposed with cloud service and data analytics as the aiding features. The read- ings are captured by mobile phone and this acts as the graphical user interface to get the status of patient's health. Also, the necessary information is shared to the doctors and authorized personnel through a web interface. Henceforth, the patient's health condition could be observed from the remote diagnosis center. The experimental results illustrate that the proposed model can continuously observe the physiologic parameters and suitably report the health status of user. © 2018 Authors.

International Journal of Engineering and Technology(UAE)

Application of data analytics for iot based healthcare system

Today, the safety of women and children has become the preeminent issue of our society. The number of incidents and victims is rising with each passing day. This paper proposes a model which will help ensure the safety of women and children. Since the person in danger will not be able to respond competently in critical situations, the need for an emergency device which easily takes input in case of emergency is the primary motive behind this paper. This paper proposes to build a prototype that can be easily carried and worn, that takes input from distress button and sends the fetched location on the cloud and notifies the respective parents regarding the same. The wearable device will have a GPS module for fetching the coordinates after the panic button is pressed. The device works with Raspberry Pi for running the GPS and updating the cloud database. The android application with the parents will fetch the stored coordinates from the cloud when they receive an emergency notification. This action allows the emergency contacts to view the most recent location of the person in trouble. The application gives the address, map location and routes of the person who needs help. © 2018 IEEE.

2018 6th Edition of International Conference on Wireless Networks & Embedded Systems (WECON)

ReachOut Smart Safety Device

Wireless sensor networks (WSNs) have been used widely in so many applications. It is the most efficient way to monitor the information. There areso many ways to deploy the sensors. Many problems are not identified and solved. The main challenge of WSN is energy efficiency and information security. WSN power consumption is reduced by genetic algorithm-based clustering algorithm. Information from cluster head to base station may have a lot of chances to get hacked. The most reliable way to manage energy consumption is clustering, and encryption will suit best for information security. In this paper, we explain clustering techniques and a new algorithm to encrypt the data in the network.

Asian Journal of Pharmaceutical and Clinical Research

GENETIC ALGORITHM-BASED CLUSTERING OF WIRELESS SENSOR NETWORK WITH NOVEL DATA ENCRYPTION

The low-cost Wireless Sensor Network (WSN) consists of small battery powered devices called sensors, with limited energy capacity. Once deployed, ACCESSIBILITY to any sensor node for maintenance and battery replacement is not feasible due to the spatial scattering of the nodes. This will lead to an unreliable, limited lifetime and a poor connectivity network. In this paper a novel bio-inspired cluster-based deployment algorithm is proposed for energy optimization of the WSN and ultimately to improve the network lifetime. In the cluster initialization phase, a single cluster is formed with a single cluster head at the center of the sensing terrain. The second phase is for optimum cluster formation surrounding the inner cluster, based on swarming bees and a piping technique. Each cluster member distributes its data to its corresponding cluster head and the cluster head communicates with the base station, which reduces the communication distance of each node. The simulation results show that, when compared with other clustering algorithms, the proposed algorithm can significantly reduce the number of clusters by 38% and improve the network lifetime by a factor of 1/4. © IJTech 2016.

International Journal of Technology

Bio-inspired, Cluster-based Deterministic Node Deployment in Wireless Sensor Networks

The demand for a Wireless Sensor Network (WSN) has increased enormously because of its great ability to supervise the outside world as well as due to its vast range of applications. Since these sensor nodes depend greatly on battery power and being deployed in adverse environments, substituting the battery is a tiresome job. Cluster-based routing techniques are prominent methods to extend the lifetime of wireless sensor networks. In this research, the work on energy efficient clustering approach is considered in two phases. During the cluster head selection phase, cluster heads are chosen which can stabilize the power consumption in sensor networks, by considering both the residual energy and distance of node with respect to sink. Later, during the cluster formation phase, a non-cluster head node will choose a cluster head that lies in close proximity with the center point between the sensor nodes and sink. Also, these non-cluster head nodes should be within the transmission range of the cluster head, as selected by the above method. Initially, the Low Energy Adaptive Clustering Hierarchy (LEACH) which is an eminent protocol for sensor networks is investigated. Furthermore, the same LEACH protocol is enhanced by proposing an effective cluster head election scheme as well as a new cluster formation scheme as mentioned above. Simulation results reveal that the proposed algorithm outperforms the traditional LEACH protocol in prolonging network lifetime. © IJTech 2016.

Improved Leach Algorithm for Energy Efficient Clustering of Wireless Sensor Network (WSN)

The recent advances in Wireless Sensor Network (WSN) have gained world-wide attention because of its minimized size, low cost and un- tethered nature. These sensor nodes has sensing unit to collect the data from physical environment, processing unit to process the data and communicating component to communicate sensory data to base station over wireless medium. WSN placement is very specific to applications such as environment monitoring, habitat monitoring, military applications, health monitoring, object tracking and smart grids. In most of these applications numerous sensors nodes are re- motely deployed and they are set to operate autonomously. WSN designers face many issues that arise due to limited battery, harsh environment, communica- tion failures, memory and bandwidth. Among this, limited power source is the primary issue that needs high priority attention. However, the open challenges in WSN such as network lifetime, connectivity, coverage are all proved as NP complete problems. For this reason several swarm intelligence (SI) techniques have been used to address WSN issues to bring efficient optimization solution. Compared to other communication and topology control protocol, the usage of SI techniques is very minimal among WSN researchers. This depicts that the WSN designers and developers are not fully aware of SI techniques. On the other hand soft computing experts are not completely aware of the WSN issues and the possibility of exploiting SI algorithms in WSN. This survey summaries the research challenges in WSN, introduces several SI techniques and tries to bridge the gap between WSN and SI researchers. It also describes the applica- tion of SI in WSN design and deployment, clustering, scheduling, localization, data aggregation and QoS Management. © 2015 Academic Publications, Ltd.