Authors:Nguyen Cong-Phuong Abstract:Durable and stable operations of induction motors are vital in industries. Overloading is one of some faults that can shorten the operating life of those electric machines. In our research, a single microphone isused to distinguish between full load, 10 percent overload, and 100 percent overload operations of induction motors. Three acoustic features and six classification models are evaluated to establish an overload classification system based on sound analysis. Obtained results show that this is a promising way to build a real-time and inexpensive monitoring system for induction motor overload.
Authors:Kashish Shah Sanket Rathi Rishabh Shetty Kamal Mistry Abstract:Language is the primary means of communication used by every individual. It is a tool to express greater ideas of ideas and emotions. It shapes thoughts and carries meanings. Indian Sign Language (ISL) used by the Deaf Community in India, does have linguistic constituents and structural properties. The area of computer science and linguistics, dealing with the relationship between computers and human language, is natural language processing. Through lexical analysis, syntax analysis, semantic analysis, processing discourses, pragmatic analysis, it processes the data. In determining the meaning of a sentence, it is critical to analyze the syntactic structure. In this paper, current computer sign language translators are considered and their pros and cons are identified and discussed. The general approaches followed by the systems are discussed. A new approach for construction of sign languages is proposed, thus resulting in increase in the accuracy of the system in translating the input phrases.
Authors:Bharath Reddy, Richard Fields Abstract:Bioinformatics is a fast-evolving topic today. It has useful from establishing phylogenetic trees, protein structure prediction to discovery of drugs and hence the importance of bioinformatics cannot be underestimated. Multiple sequence alignment (MSA) is the main step in performing the above tasks mentioned. Multiple sequence alignment is the science or a method where more than two sequences are arranged one above the other to find the regions of similarity between them. These regions of similarity are called ‘conserved-regions’. Over time, there are many algorithms which are developed to give a ‘good’ alignment. These developments were essential to construct phylogenetic reconstruction, protein structure and protein prediction accurately. In this paper, we will talk about the most popular multiple sequence alignment algorithms. We first begin with the definition of multiple sequence alignment. Thereafter, we shall talk about the different techniques in multiple sequence alignment along with the most popular MSA algorithms
Authors:Akileng Isaac Aruna Bhat Abstract:Hate speech is by no means always on the rise due to the high rate of remote service usage such as communication, online studies, meeting, dating, etc. With the recent outbreak of COVID-19, there has been an increase in the number of users on different social media platforms. This increase in number has brought about an increase in issues such as hate speech, among others. This paper aims to provide a detailed process of improving LSTM used for hate speech detection using knowledge distillation. The knowledge transfer is done from the more extensive network (teacher) to the smaller student network. The teacher has trained for five entire epochs to output accuracy of 76.8%, the stu-dent network trained from the teacher network for three whole epochs attained an accuracy of 82.6%. Another student model cloned and trained from scratch for three entire epochs instead of the teacher network achieves an accuracy of 75.4%.
Authors:ChukwuNonso H. Nwokoye Vincent O. S. Okeke Paul Roseline Ethelbert Okoronkwo Abstract:In recent times, African nations have been mostly absent in dis-cussions concerning Artificial Intelligence (AI)-powered driverless cars. Additionally, it was also discovered that several global surveys and other studies on driverless car acceptance, popularity and confidence excluded Africa. This is in the light of its immense benefits which include the re-duction of road accidents, an effectual car-sharing and transport structure and accurate navigation with less consideration of distractions. There-fore, we examined the challenges and risks attendant to the deployment of self-driving cars in a developing region such as Africa. Several chal-lenges were identified, they include lack of needed infrastructure, ab-sence of law and order, cost, absence of image detection and recognition projects, absence of practical artificial intelligence course-ware, need for an advanced AI-based algorithm, weak legal framework and other ethical issues, criminalization , security and privacy, high tendency to cause more unemployment. The paper highlights several risks attendant to such forms of advancement in Africa.
Authors:Siddharth Srivastava Yash Chaudhari Yash Damania Parul Jadhav Abstract:Automated Image Captioning involves understanding the semantic information of an image and expressing it in natural language. Among the many approaches proposed, deep learning based techniques have achieved state-of-the-art results in solving this problem. In this paper, three primary, distinct deep learning based approaches to solve this problem are introduced and compared: encoder-decoder frameworks, neuroevolution, and attention based approaches. This paper covers their mechanisms and their performance, and highlights where they di er from each other. To conclude, the results of these approaches on benchmark dataset and metrics are presented.
Incoming Master's Student at Cornell Tech, NY, US. Undergraduate Student in CS at MIT-WPU, Pune, India. Interested in AI & Physics (Presenting "Deep Learning Techniques for Automated Image Captioning").
Thursday April 15, 2021 3:00pm - 3:15pm BST
Virtual Room BLas Vegas, Nevada, USA
Authors:Maria Krommyda, Nikos Mitro, Angelos Amditis Abstract:The building materials of Cultural Heritage monuments are subjected to continuous degradation throughout the years, mainly due to their exposure to harsh and unexpected weather phenomena related to the Climate Change. The speci c climatic conditions at their vicinity, especially when there are local peculiarities such as onshore breeze, are of crucial importance for studying the deterioration rate and the identi ca-tion of proper mitigation actions. Generalized models that are based onclimate data can provide an insight on the deterioration but fail to offer a deeper understanding of this phenomenon. To this end, in the context of the EU funded HYPERION project a distributed smart sensor network will be deployed at the Cultural Heritage monuments in four study areas as the solution to this problem. The platform includes smart IoT devices designed to provide environmental measurements close to monuments, a middle-ware to facilitate the communication and a visualization platform where the collected information is presented.
Authors:Sreesankar Ajayan Gouri H Radhika Menon Georg Gutjahr Prema Nedungadi Abstract:The Amrita Awareness Ambassador (AAA) program trains local youth in tribal areas to act as ambassadors to help prevent sub- stance abuse. With the support of the Ministry of Tribal A airs, Govt. Of India, the program is active in 13 states. The present paper analyzes data from the program from 15 villages in Chhattisgarh and Jharkhand. 609 children from these areas participated in a survey prior to the pro- gram and a follow-up survey six month later. Costs of conducting the programs in 2019 and 2020 in these villages are presented. E ectiveness is measured as bringing long-term awareness on substance-abuse to a child who did not previously possess such awareness. The program is found to be cost-e ective for a willingness to pay of INR 2770.
Authors:Shouherdho Banerjee Akash Fardeen Mahbub Rashedul Islam Sayed Abdul Kadir Al-Nahiun Raja Rashidul Hasan Md. Abdur Rahman Abstract:Lung Cancer and Tumors have been a common cause of death in recent times. Early identification of that kind of cancer is challenging due to its small size. Among current techniques, Microwave Imaging (MI) has been one of the efficient methods for diagnosing lung cancer. In this research work, a Microstrip Patch Antenna has been designed and simulated in the CST Studio Suite Software for MI to diagnose lung cancer using the FR-4 (Lossy) substrate material and a resonant frequency of 2.3 GHz (5G-Band), that ranges from 1.5 GHz to 3 GHz with a 60.46*78.73*1.7 mm3 in dimensions. A Lung Phantom has also been cre-ated involving lung tissues, pleura, muscles, fats, and skin. After that, a tumor of 5mm size was placed on that lung phantom both with and without the tumor for measuring the Antenna's output performance. After implanting the Antenna in the lung phantom with and without tumor, a Return Loss (S1,1) of -47.30 dB and -48.93 dB were respectively obtained. This research also introduces other ob-tained performance parameters such as Radiation Efficiency, Polar Radiation, Directivity (3D), Gain, SAR, and others, illustrating that the antenna model is a safer choice for detecting lung cancer on time
Authors:Deepali A. Mahajan. Namrata Mahender C. Abstract:In today’s digital scenario the use of social networking site has become popular in students. They are feeling comfortable and convenient while using this technology for different purposes. Many times this excess turns to be dangerous for the youth. As they get distracted from their studies and works. We try to find the behaviour of students towards the use of WhatsApp in our studies. The study conducted on the 98 college and university students that are of the undergraduate and postgraduate level students. We have distributed self-generated question-naires among the students and recorded the responses. We found that the regular but moderate use of WhatsApp. The use of WhatsApp keeps students happy and feeling excited. WhatsApp addiction is not found in the students.
Authors:Sujatha Krishnananthan Puvanendran Sanjeeth Rukshani Puvanendran Abstract:This Study focuses on the prediction of the Diabetic Patients through the sugar levels. The Dataset is analyzed using the data mining techniques such as feature extraction, associate rule mining and classification. The Fast Blood Sugar (FBS) and Post-Prandial Blood Sugar (PPBS) sugar levels are selected as the important features, identification of a rule depending on the selected feature is identified and the performance metric for three classifiers is analyzed based on the selected attributes and choose the classifier with high accuracy. Classification algorithms like Random forest, Decision Tree (J48), and Naïve-Bayes are utilized to identify the patients with diabetes disease. The performance of these tech-niques is considered using the factors relating the accuracy from the applied tech-niques. The Accuracy is seeming to be higher for Naïve Bayes. The outcomes acquired demonstrated that Naïve Bayes outflanks from different strategies with most noteworthy precision of 74.8%.
Authors:Saksham Gupta, Kashyap Joshi Abstract:A powerful astute ground vehicle is planned which can perform different errands like recognizing objects and evading the articles self-rulingly and furthermore it can distinguish paths to move itself whenever needed to ride on a street. The vehicle comprises sensors and camera to identify objects to evade them and it can likewise move on a predefined way because of the utilization of gps framework on the vehicle. It very well may be utilized in the midst of catastrophe to help individuals by checking and identifying presence of harmed individuals in a huge region and giving them fundamentals. It can likewise be utilized as a conveyance vehicle in the midst of pandemic like COVID-19 where contactless conveyance of fundamental merchandise should be finished.
Authors:Praveen Manoj Singh, Sathidevi P.S. Abstract:A machine learning based technique for the detection of unipolar and bipolar depression disorders is developed and implemented in this paper. Unipolar depression and bipolar depression share almost similar clinical symptom pro le and hence the diagnosis of the type of depression is a great challenge. Disturbances in motoric activity may be a useful way to detect pathological mental states. Hence, a unique motor activity database (Depresjon) which is collected from patients su ering from unipolar depression, bipolar depression and healthy subjects is em- ployed in this work. Performance of the proposed method is evaluated using sensitivity and speci city metrics. Results show a sensitivity value of 0.98 and speci city value of 0.88 for unipolar depression and sensitivity value of 0.88 and speci city of 0.98 for bipolar depression.
Authors:Shivom Keshary Ganeshaperumal Dharmaraj Subathra Balasubramanian Seshadhri Srinivasan Abstract:From the last decade Internet of Things(IoT) brought tremen- dous changes for industries and Healthcare. It provides the Smart digital platform for remote healthcare monitoring. With the help of IoT technology, patients can monitor and record their vital sign parameters such as Body Temperature, Heart Rate, and Oxygen Saturation level. In this paper, the Heart Rate and Oxygen Saturation level monitored using the MAX30100 pulse oximeter sensor, and a DS18B20 temperature sensor is used for body temperature monitoring. The NodeMCU ESP8266 IoT platform device with Arduino software used as Gateway for sending sensor data to the cloud and mobile app. For visualization and insights of data, sensor data send to ubidots cloud and the Blynk mobile app. This work contributes to providing personalized vital signs monitoring and sending real-time patient health data remotely to the Doctors and caregivers.E36
Authors:Md. Sayeduzzaman Md. Samiul Islam Borno Md. Hasibul Islam Sujan Howlader Abstract:Internet of Things (IoT) is the connection between devices by the use of the internet. IoT conducts logical operations by which data is transmitted and processed in the cloud. Structural improvement and modernization are required. This paper focuses on advancing IoT technology by creating an automated sani-tization device that could have been used to protect health by sanitizing the home's inner surface. Smart homes with auto sanitization can be a blueprint for reducing the transmission rate of COVID-19. It can also be an example of a cost-effective and time-saving approach to have a clean, comfortable, and happy liv-ing life. Smart home with an auto sanitization system had proposed a fog disin-fection machine, NodeMCU, Arduino Uno, and various sensors to create the pro-totype. This device can be used on the internet via a mobile application from anywhere in the world.
Authors:Divyashikha Sethia Ritik Aggarwal Saksham Bhayana Sanchit Mehta Abstract:Portable IoT devices in uence everyday usage and communi- cation, such as a programmable thermostat and smart lights in a smart home. Advanced technology can assist in remote management and se- lective authorized access by di erent users. The devices must also be secure from malicious users and have minimal overheads for revocation while providing uninterrupted access to valid users. Ciphertext-Policy Attribute-Based Encryption (CP-ABE) scheme can provide ne-grained, selective access to these portable IoT-based devices. This paper proposes a novel RSA-based CP-ABE Scheme with Scalable Revocation (RCSR) scheme for IoT and Smartcard devices, which extends an existing RSA-based CP-ABE scheme for scalable user revocation and a x for a security breach for the collusion of keys. The proposed RCSR scheme is lightweight, with constant computation time and storage space for both keys and ciphertext. It provides e cient user revocation with uninter- rupted access to valid users. Detailed quantitative performance comparison indicates that the proposed RCSR scheme performs better than the previous schemes with acceptable overheads.
Authors:Xuehong Wang Chunling Tu Pius Adewale Owolawi Abstract:The accuracy of classification is a very important step in strawberry maturity estimation. In order to improve the accuracy, a backpropagation (BP) Neural Network is employed with a structure 3-6-3, in which there are three inputs: the mean and the variance of the Hue component, and the red area ratio of strawberry and three output layers: maturity probabilities of unmatured, almost matured and matured. The mean and the variance of the hue component in the hue-saturation-intensity (HSI) colour space are extracted from the strawberry pictures. Besides, the red area proportion of the strawberry is analysed. The Neural Network is trained with a dataset that we captured. These imputes are characteristic number of strawberries. Which are simpler and faster than an image as imputes. The results show the accuracy of the proposed method is above 96%, +E26which is high compared with the existing methods.
Authors:Kuljeet Singh Amit Mahajan Vibhakar Mansotra Abstract:Port scanning is a frequently used service in the internet with its usage done by both network defenders as well as attackers. Detecting Port scan attempts is of vital importance to prevent attackers from gaining sensitive information about the network. In this paper, a deep learning (DL) based model using Convolutional Neural Network (CNN) is built for classification of Port scan attempts in the CICIDS2017 dataset. Feature selection techniques like Recursive Feature Elimination (RFE) and Fisher score have been incorporated with DL models for selecting best features. Different values for optimal number of features are experimented upon to study the effect of dimensionality reduction. Experimental results have shown satisfactory performance in each of the metrics for all the models trained and evaluated.
Authors:D. Tejesh Reddy Shamini S Venithraa Ganesan Anisha Radhakrishnan Vineeth Ajith Abstract:Out of all the basic needs of humans, water is the most essential for livelihood. It's clearly known that water is not distributed evenly in the earth's surface.Out of total only 3% of water is fresh water and the remaining 97 percent of water is not t to drink.Of the total fresh water, 69 percent of them resides in glacier, 30 percent of underground water, and less than 1% of water is situated in rivers, lakes, and swamps.Apparently all the people should have access to this water appropriately. There are many places where community members still struggle a lot to access the drinking water. This paper deals with such a situation where people don't have access to clean drinking water.To tackle this situation ,feasible puri cation methods to solve the problem are proposed.
Authors:Poojan Panchal Vignesh Charan Raman Trupti Baraskar Shambhavi Sinha Swaraj Purohit Jaynam Modi Abstract:In the field of remote sensing satellite imagery, malfunctions in the available raw data are prominent. Especially in Short Wave Infra-Red (SWIR) detectors used in satellite imaging cameras, which suffer from dropouts in pixel and line direction in raw data. With the recent development in Generative Ad-versarial Networks and its vast application in inpainting the missing data, the possibility to predict and fill in the missing data accurately with contextual at-tention has become prevalent. This paper presents SN-GANs (SN-Generative Adversarial Networks) which is two-stage architecture, and it is based on the concept of feedforward neural networks with contextual attention layers. While reconstructing the corrupted part of the images the model takes surrounding pixels into consideration. Moreover, this architecture is adept enough to fill in the multiple lines and pixel dropouts efficiently even in super-resolution satel-lite images. The available traditional methods fail to address the loss of data that incurs while inpainting a 16-bit raw image because they are effective enough for 8-bit RGB images. SN-GANs has effectively resolved this issue with a lossless image inpainting method for 16-bit satellite images as it retains the features of non-corrupted data. The performance of the model is evaluated using similarity metrics like Structural Similarity Index Measure (SSIM), Peak Signal-to-Noise ratio (PSNR) and Mean Squared Error (MSE).
Authors:Md. Hasibul Islam Khadija Yeasmin Fariya Md. Taslim Hossain Tanim Touhidul Islam Talukder Nafiz Ahmed Chisty Abstract:The new age cannot be imagined without advanced technology. In many aspects of our lives, automated systems have taken over traditional sys-tems. IoT plays a significant role in all automated devices. Road accidents have become a major concern today. In several occasions, persons have died because of not receiving emergency treatment services following an injury. This paper mentions a cost-effective IoT-based innovative system to track the accident from the authority's control room. The system can detect road accidents and notify the concerned authority by sending the location and car number. Additionally, an emergency push button with face detection has been incorporated to help anyone in distress. The self-powered system can remotely monitor the street light and increase or decrease its light intensity. Deep learning has been used for imple-mentation, and the system can be operated by a mobile application from any-where at any time, and the data will be updated periodically on the server. It is believed that this project will have a massive impact on society and will bring positive results.
Student, American International University-Bangladesh
Hello, I am Md. Hasibul Islam. I am from Bangladesh and recently completed my graduation from American International University-Bangladesh. I am involved in research and publications. Now I am trying to build up my career with this research. My research interested areas are Electrical... Read More →
Thursday April 15, 2021 5:45pm - 6:00pm BST
Virtual Room ALas Vegas, Nevada, USA
Authors:Rajesh Saturi P. Prem Chand Abstract:Histology is the study of tissues examining under a microscope to identify the severity of disease. it became very critical in biomedical research and clinical practice. Processing tissues from histopathological images has become now fully computerized, pathologist observe digital pictures on a computer rather than on microscope in order to predict the seriousness of cancer. Routine analysis of tissues selection will be very difficult, manual task can be done only by trained pathologists at a huge cost. Hence we need some supervised algorithms to identify stage of cancer and the seriousness of cancer so that it helps in diagnosis of disease. It is a very difficult task, because of cell nuclei are very small, irregular and contains significant noise. Therefore, an optimized segmentation method was suggested for nuclei recognition in histopathology breast cancer images. For experimental analysis the dataset is collected from Break His database. And next, pre-processing is done by applying image normalization (contrast stretching) which improves image quality. After enhancing the quality of image, an iterative clustering algorithm is employed to select the super pixels and later a novel variant of Particle Swarm Optimizer (PSO)-Gray Wolf Optimizer (GWO) search algorithm is employed to select the optimal cluster centroids. The purpose of this proposed system is to perform segmentation and find optimized cluster centroids for automatic nuclei cell recognition. Hence the proposed work achieved good results compared to the previous work. Further we can separate the overlapping cells from segmented image and apply supervised learning to predict the stage of cancer.
Authors: Pranjali Deshpande, Sunita Jahirabadkar Abstract: Text summarization is an important application of Natural Language Processing (NLP). A huge amount of data is generated everyday through the internet, newspapers, etc. Quick understanding of the documents help reader to save time, retain interest in the reading and provides the clarity of the content. Text summarization facilitates this by two approaches - Extractive and Abstractive. Where extractive approach retains the key phrases and key sentences in the document, abstractive approach focuses on generation of new summary sentences by understanding the crux of document. Summary generation becomes more challenging in case of low resource language documents, as low resource documents lack the large corpora. This paper intends to analyze and compare the techniques used for the abstractive summarization of low resource languages.
Authors:Ceronmani Sharmila V Mohamed Aslam H Mohamed Riswan M Abstract:Connected and autonomous vehicle industry is fostering advancements in technologies such as Artificial Intelligence, machine-learning, Internet of Things and Bigdata applications. Connected and autonomous vehicles rely on a variety of electronics, sensors and computer systems. They require strong cyber security mitigations to ensure that these systems work reliable and to reduce security risks. While the possibilities for integrating devices and vehicles seem endless, new threats and dangers arise every day. Due to their connectivity, there are also security risks to the networks they are connected to, such as road network sensors, electrical infrastructure or vehicle control features and so on. Systems should be created in order to reduce all the possible threats and vulnerabilities. In the study, we will analyse the various cyber-security vulnerabilities and threats which are possible in autonomous vehicles by constructing graph from observing harm propagation of the cyber threats and propose a mitigation model for the same.
Authors:Rahmat Yasirandi, Wulandari, Paulus Berliz Sitohang Abstract:Piots Tanah is a plan-based sensing system that developed to meet the learning needs at the vocational-high school level (especially the majors related to agriculture). This system claims to be able to sense several parameters related to the environmental conditions of a plant. Through the Technology Acceptance Model (TAM), this study looks at the level of readiness for the adoption of the proposed technology. with the addition of the System Features of Piots Tanah variable at TAM, there are 4 hypotheses on this research. The reason for this variable isadded, because it is clear that the system features o ered have an impact on user intention.Furthermore, each instrument variable from the questionnaire has a question guaranteed validity. Seen from the Cron-bach Alpha' result, each value of Alpha of all variables is greater than 0.7 and the average value is 0.736. Then the results of the questionnaire distributed to users have been eligible for hypothesis testing using regression analysis.H1, H2, H3, and H4 are accepted, because p for each hypothesis is 0.05. With the highest value of owned by Hypothesis 4 (there are positive e ect on Behavioral Intention from Attitude). In fact, R2 also shows a declining value (R2 on attitude 0.38 and R2 on behavioral intention = 0.43). Declare that the adoption of the model used is appropriate. because the value of impact (R2) of the variables that a ect these 2 variables is quite high.
Authors:K. Sujatha G. Nalinashini B. A. Ganesan N. Jayachitra RengammalSankari A. Kalaivani S. Sendilvelan Abstract:Supply of dry air is of important use in industries because the products manufactured need to be moisture free. The dry air can be generated by heating the air using a heater element through which the power needs to be supplied as input. This drying process needs to be controlled in order to obtain a desired output temperature. Dynamic modeling of the dryer is of great importance, because by controlling input power the output temperature of dry air is varied so that it is used for drying of sugarcane at sugar factory. The mathematical model of the sugar processing plant is fitted to its corresponding data is capable to tracking variations in the controller output and also can tackle the external disturbances. Simulation studies are carried out compare the drying process output with conventional PI feedback control and the novel Model Based Predictive Control (MPC).The proposed control algorithm is of advanced type and is used to control the temperature of drying process by varying the input power. The drying process has a ramp shaped output which denotes that the MPC approach is beneficial as compared with the conventional methods.
Authors:Bernard H. Ugalde, Albert A. Vinluan, Jennifer T. Carpio Abstract:Proper travel planning is an important consideration for people with ambulant disabilities for them to be able to move around safely with agility. Thus, determining the appropriate route after considering a variety of converg-ing factors is a crucial task. The purpose of this paper is to analyze how the ap-plication of Mathematical Risk Modeling and Analytical Hierarchy Process op-timizes the graph representation for people with reduced mobility. This paper first attempted to identify the highly relevant factors influencing the selection of the best route in the Central Business District of Baguio. Then, it presented a novel model that uses scientific calculations and subjective judgments in the decision-making process. The proposed model strengthens the capacity of peo-ple with ambulant disabilities to handle ambiguous navigation criteria with some time and safety benefits.
Authors:Priti Shahane Ujwala Kshirsagar Abstract:Network-on-Chip (NoC) has been proposed as an evolving solution for the scalability and performance demands of the next-generation System-on-Chip (SoC). NoC provides a way out for interconnection issues of bus system in SoC, where dense numbers of modules are integrated for better performance on a single silicon chip. The router is a most important device that significantly affects NoC architecture performance. Buffers in the router provide high throughput at the cost of an increase in significant area and energy. Bufferless deflection routing offers a solution for improving energy efficiency, but it may lead to an increase in latency due to unnecessary hopping of data packets. Use of a small side buffer helps to leverage the area and latency of the router design. Another aspect of router design is an efficient design of scheduler. Iterative Serial in Line Protocol (iSLIP) scheduler with programmable priority encoder provides starvation-free scheduling. Here, we have proposed a mesh topology for NoC router design with small side buffers in the input block and a starvation free iSLIP scheduler. The proposed router design is implemented for 4x4 mesh topology on the FPGA board and the results are compared with other FPGA and ASIC based NoC router architectures. The proposed design has resulted in a 12% reduction in logic resources and a 2.5 times better operating frequency than the CONNECT NoC router.
Authors:Denis Stanescu, Angela Digulescu, Cornel Ioana, Alexandru Serbanescu Abstract:Power transmission and distribution networks currently face an in-tensive use and climate change challenges. In such networks, the occurrence of defaults (caused by the cable ageing process, the structure and network capaci-ty, voltage levels, and related environmental conditions) is inevitable. The diffi-culties of predictive maintenance of power grids are related to the large spread of electrical infrastructures and the definition of early warning indicators. Such indicators are the partial discharge activities which are very informative about the rising insulation problems of electrical materials. The monitoring of such phenomena is nowadays an important field in the power transmission and dis-tribution systems. The purpose of this paper is to study different techniques for the detection of transient signals in the context of monitoring partial discharges in electrical networks. Among the various techniques developed and used for the detection of partial discharges, we will focus here on four classes of meth-ods; two of them are encountered in existing products, whereas two new tech-niques are based on recent concepts: compressive sensing and phase-diagram based analysis. The comparative study concerns the evaluation of the probabil-ity of detection versus false alarm ratio of real signals measured on a reduced scale experimental facility. Thanks to this experimental facility typical for real life configurations, we are able to add to the partial discharge signals perturba-tions such as load signals, reflections but also the effect of propagation effects in a real electrical cable which allow us to highlight the performances of each method.
Authors:Payal Bose Prof. Samir K. Bandhopadhyay Abstract:Detecting human and human motion through a video surveillance sys-tem is an important task. It helps in various fields, like crowd analysis, identifying abnormalities in human behavior in public places, identifying a person and their gender, etc. To detect a human and its motion in a video the first step is to detect the moving object correctly. Background subtraction method is an efficient way to detect the moving object efficiently on the foreground frame. Object detection and extraction could be performed in various ways. Once the object is found then a classification method is applied to recognize the object and its motion. In this paper, first the background subtraction method is applied to detect the person in a video input file. Then this method is applied to track the human’s motion through the entire video. In the second step, the Histogram Oriented Gradient Method (HOG) is used to extract the features from the input video file. And in the last step Support Vector Machine (SVM) is used to classify the detected per-son’s identity and its motion. In this experiment two different multi-class classi-fiers are used, SVM and Decision-Tree (DT) to compare the performance of the classification models. Finally, a detailed comparison of each person and the mo-tion class is performed to compare the classification rate of the two classifiers.
Authors:Jia ni. Zhou Kwang ho. Baek Tae soo. Yun Abstract:In this paper, we proposed that at certain camera distances, Houdini and Unreal can achieve approximations that result in shorter final rendering times. Based on the optimal polycount value of the interleaving mesh from Houdini to Engine, the focus is on whether the distance of the camera has an ef-fect on the similarity of the final results of the two software. we used the ren-dering images of ocean wave effects at different camera distances as the object of comparison, using the similarity table derived from the image comparison system, so as to derive the highest similarity when the ocean wave effects in Houdini are imported into Engine, the camera distance value field, and to re-store the Houdini rendering effect in the Engine as far as possible. In subse-quent research, a solution will be proposed to the problem of lack of accuracy when close-up scenes in UE4, testing the effect of materials on overall similari-ty.
Authors:Pramod B. Deshmukh Kanchan Lata Kashyap Abstract:Breast cancer (BrC) is among the biggest causes of death for women in today's world. Manual identification of BrC cells is a tedious assignment that requires human blunder and, as a result, computer-aided detection (CAD) assisted components are used to achieve higher outcomes as contrasted and manual obsessive discovery frameworks. The breast-image classification mission advanced engineering of natural image classification techniques and methods of artificial intelligence have been used to a large extent. Advanced histopathology image examination is a promising way to deal with accurate and cost-effective cancer disease determination. The consideration of computerized picture arrangement gives a subsequent assessment and saving time for the specialist and the doctors. There are not many survey papers accessible that incorporate a complete outline of BrC image classification methods, selection strategies, feature extraction, classification estimating rules, and results of image classification, despite the numerous publications on breast image classification. The goal of this review is to evaluate, identify and resolve current developments in the detection of human malignancy using AI, machine learning (ML) and deep learning (DL) methods for BrC. The study illustrates how disease analysis as well as remedial action is supported by the use of ML and DL techniques. To improve analytical accuracy and reduce emotional error, the CAD system assists pathologists in classifying medical images into normal, benign, and malignant cells.
Authors:Yamuna K. Kalpesh Gupta Sradha D. Prabhu Navaneeth Krishna S. Sai Jyothi Jayasree Narayanan Sani S. Abstract:Water is an ineluctable integral ingredient for the survival of all living beings. During the next century, more than a quarter of the world's population of developing countries is projected to face ex- treme water scarcity. This paper shows the drinking water shortage and methods to purify contaminated water in Krishnarampur, a village in West Bengal, India. An ethnographic research approach was used for an in-depth study of problem detection. Dumping of generated waste, and activities like washing clothes, washing utensils, and bathing were deduced as the factors responsible for contamination of the local water sources lakes, ponds, and rivers. Only certain water resources of the village were found to be clean and t for use, but not drinkable due to dissolved contaminants. This study discusses technological propositions to improve the quality of pond water for the village community.
Authors:Vandana Chaudhari Husain H. Dawoodi Manoj P. Patil Abstract:Plant diseases are one of the main causes to reduce the quality and quantity of crop yield. Traditional expert eye necked opinion method is used to recognize the crop diseases. In rural areas getting experts, opinion is a more te-dious job. The automated system is required to recognize the disease accurately and quickly. In automated systems, feature extraction played an important role. Many feature extraction techniques are suggested for the classification of plant diseases. According to the literature, the Local Binary Pattern (LBP) method has good object recognition and classification performance. In our work, we have compared a study of extracted features based on the LBP and Gray Level Cooccurrence Matrix (GLCM) feature with color features for the recognition of banana plant leaf diseases. The experiments were performed using different ba-nana leaf diseases like banana yellow sigatoka, bunchy top, cucumber mosaic virus, and healthy leaves. The result shows that the features extracted using LBP are more accurate as compared to GLCM when compared with the SVM classi-fier. KNN gives more accurate results with GLCM.
Authors:Olabode Idowu-Bismark Francis Idacheba Aderemi A. Atayero Walter Janusz Caitlyn Harling Abstract:The understanding of root mean square delay spread (RMS DS) is important in wireless communication for mitigating intersymbol interference. Therefore modelling the RMS DS will help in understanding its characteristics and useful in designing wireless signal receivers. In this work, our dataset were gotten from deterministic predictions based on simulations using the wireless Insite X3D ray tracing engine which uses site-specific geographic databases of Broad Street Lagos Island Nigeria for terrain, buildings and foliage. We characterize the delay spread in a street canyon and a high-rise environment and provide a model for the two scenarios. The model statistics were presented. The achievable channel capacity for the two scenarios were obtained and modelled. Cross correlation coefficients between the RMS DS and channel capacity were used to explain the effect of RMS DS on capacity as height of the building and street distance increases.
Authors:Prapti Kachhia Dr. Dushyantsinh Rathod Abstract:Health Care field has an immense measure of information, for processing those information certain methods are utilized. In healthcare industry, data mining plays an important part of predicting diseases. For detecting a disease numerous of tests should be required from the patient. Kidney disease is the main organ in a human body. In any case, presently a-days Chronic Kidney Disease (CKD) is the most widely recognized issue for people. Today, a few people die because of Chronic Kidney Disease. This disease is the most well-known and a serious disease in the world. The progressive loss of capacity of a kidney is additionally called Chronic Kidney Disease. Machine Learning is a promising methodology which helps in early diagnosis of disease and might help the professionals in decision making for diagnosis. This Dissertation analyzes data mining techniques which can be utilized for predicting like Kidney diseases. we calculate accuracy of machine learning algorithms for predicting Kidney disease, for these algorithms are Naïve Bayesian, LR, Optimized XGB Algorithm by using dataset for training and testing and we found to better accuracy in that system. The point of this investigation is to build up a framework which may predict the Kidney disease hazard level of a patient with a superior precision.
Authors:Luis Cuautle Gutiérrez, José de Jesús Cordero Guridi, Johannes Carrillo Aguilar, Eduardo Lebano Pérez Abstract:The present work consists of the creation of a virtual driving experi-ence of an all-terrain vehicle that includes the use of an industrial robot in order to promote engineering careers at a private Mexican university. In its develop-ment, a product design and development methodology was used that consisted of virtualizing the route, in this case an off-road track, using computer-aided design and the Unity platform. The robot programming consisted of simulating seven physical movements of the path. For the validation of the experience, a questionnaire was created with 13 items that covered visual aspects such as the recreation of the vehicle's interior, the route such as curves, ascents and de-scents, and the general experience in terms of duration, sounds and similarity. to a real environment. It was applied to a non-probabilistic sample for the conven-ience of undergraduate students and was validated based on its Cronbach alpha value. The findings show that the participant's experience and physical integrity are the strengths achieved with this development, a situation that allows its use for the stated purposes.
Authors:V.Ceronmani Sharmila Nandhakumar N P Akkash Babu N S Bharath Kumar S Abstract:The Industrial Fourth revolution (industry 4.0) is achieved through deploying IOT (Internet of Things) Device over Industrial Area due to communication latency and geographical distribution of Cloud centric IOT devices in industrial environment fails to acquire quality of service as per the requirement. An effective solution for this problem is provided through fog computing which uses edge resources to execute the applications which minimizes service delivery time of various applications by coordinating fog nodes with the IOT devices. The node which visits all other node within a virtual network is selected as a first node instead of selecting node that has maximum density. Then further other nodes are selected based on the density according to the applications need. Energy of the nodes is utilized efficiently in terms of average hope and size using Connected Domination Set (CDS).
Authors:Zheng Li, Mauricio Pradena Miquel, Pedro Pinacho-Davidson Abstract:Despite the fact that outside is becoming the frontier of indoor workplaces, a large amount of real-world work like road construction has to be done by outdoor human activities in open areas. Given the promise of the smart workplace in various aspects including productivity and safety, we decided to employ smart workplace technologies for a collaborative outdoor project both to improve the work efficiency and to reduce the worker injuries. Nevertheless, our trials on smart workplace implementation have encountered a few problems ranging from the theoretical confusion among different stakeholders, to the technical difficulties in extending underground devices’ lifespan. This triggers our rethinking of and discussions about “smart workplace”. Eventually, considering the unique characteristics of outdoor work (e.g., more sophisticated workflows and more safety-related situations than office work), we argue that “safety-centric and smart outdoor workplace” deserves dedicated research attentions and efforts under the umbrella discipline of smart environment. In addition, the identified technical challenges can in turn drive different research dimensions of such a distinguishing topic.
Authors:Anita Ramachandran Atul Kumar Pant Anupama Karuppiah Abstract:Sleep apnea is a sleep disorder in which a sleeping person’s breathing is disturbed. Subjects suffering from sleep apnea undergo periods of no or shallow breathing during their sleep. Sleep apnea may lead to severe issues such as diabetes, cardio vascular problems, hypertension, neurological issues and liver problems. Because of the global prevalence of sleep apnea as well as the direct and indirect long term problems it brings about, it is important to diagnose and treat this condition. Sleep apnea is detected clinically by the Polysomnography (PSG) test which measures various biomedical parameters such as electrocardiogram (ECG), electroencephalogram (EEG) and oxygen saturation (SPO2) over a full night’s sleep. The application of machine learning to detect sleep apnea from these parameters has gained ground in the recent past because of its ability to learn from the training datasets and generalize well to make predictions on new data. In this paper, we look at the performance of 6 machine learning classifiers–k-nearest neighbors (kNN), Artificial Neural Networks (ANN), Random Forest, XGBoost and Support Vector Machine (SVM) in their ability to detect apneaic events. The study is based on a datasets with ECG signals.
Authors:James Addy, Michael Cameron, Md Hasan, Augustine Ukpebor, Kamal Ali, Ali Abu-El Humos Abstract:The Underwater Magnetic Release (UMR) is a device developed by Jackson State University (JSU) MBRACE [8] research project team to allow for the deployment and retrieval of an underwater platform without the need for a float marking its location, minimizing its exposure to vandalism. The device is programmed prior to deployment with the expected retrieval date and time. The retrieval time, accurate to the second, can be set as far ahead as a year in the future. At the programmed retrieval time, a float attached to the platform is re-leased to the surface marking the platform’s location for retrieval. This UMR is ideal for underwater experiments where equipment needs to stay submerged for prolonged periods of time without having a surface location marker that might attract unwanted attention. This system is particularly beneficial in oyster re-search by keeping the oyster cages and underwater scientific equipment safe from vandals. The system can also be of use by trap fishermen eliminating the need for a surface float until it is time to retrieve the trap. The UMR described in this paper is one that is programmable through a smart device application to operate for set periods of as long as a year.
Authors:Manish Pant A. K. Shukla Sanjay Kumar Abstract:We purpose a novel intuitionistic fuzzy time series (IFTS) forecasting method to forecast death due to COVID-19 in India in this study. The purpose method uses an enhanced conversion method of Singh et.al [1] and novel definition of Cartesian product of intuitionistic fuzzy sets (IFS). Performance and accuracy of purposed forecasting method is verified using correlation coefficient (R), coefficient of determination (R2), statistical parameters (perfor-mance parameter (PP), tracking signal (TS)) and error measures. We have also optimized the length of interval for purposed forecasting method using a MATLAB function to minimize the forecasting error in forecasted outputs.
Authors:Bilgehan Gurunlu, Serkan Ozturk Abstract:Although image processing and forensic computing are different fields, they have been involved in the same computer science research areas such as image forgery detection, in recent years. Image forgery detection is a new branch of image processing due to increased image manipulation tools. Thus, we proposed a new block-based image forgery detection method within this frame-work. In this research, we applied the latest and easiest application feature ex-traction method used in a new iris recognition system, called rotation invariant neighborhood based binary pattern, on the block-based image forgery detection system. To the best of our knowledge, this is the first work that applies to block-based copy-move forgery detection systems. The proposed method has been eval-uated for different block sizes on a well-known image database (CoMoFod) in the literature. Experimental studies showed that our method forgery detection ac-curacy rate incentive results are higher than the state-of-the-art block-based for-gery detection methods.
Authors:Anamika Chauhan Rishabh Lokesh N. Shankar Pratham Mittal Abstract:Blockchain is the backbone of digital cryptocurrency systems, and it eliminates the need for a central authority in a decentralized network. Blockchain is a system that allows the sharing of information securely and transparently in a peer-to-peer connected decentralized network consisting of untrustable users. Blockchain technology ensures security, equality, and fairness of the system by following a well-defined set of rules known as a consensus protocol. These consensus protocols are at the core of blockchain technology and secure the network from various attacks and frauds. It is nearly impossible to breach a system following these protocols. There are various consensus protocols, each with its advantages and disadvantages. In this paper, we have discussed various consensus protocols in terms of performance, fairness, resource requirements, and security breaches. We have discussed the working of typical blockchain technology by describing major concepts in the implementation of bitcoin technology. We also discussed the most common security breaches in decentralized systems.
Authors:Mayssa Bessifi Ahlem Ben Younes Leila Ben Ayed Abstract:Business Process Modeling Notation (BPMN) has acquired increas-ing relevance in business process modeling. However, it suffers, as all semi-formal languages, from the lack of rigorous semantic which hinder its full adop-tion, and make impossible the verification of relevant properties. In this paper; a model-driven approach for the verification of BPMN2.0 models using the Event B method is proposed. When the complexity of the target system increases, the traceability of Event B model elements back to the original BPMN model be-comes a tedious task and failures to verify properties can be difficult to under-stand. For this reason, the proposed approach is adapted to keep trace of the transformation process in the format of a serialized file. It is automated through mapping rules implemented in a first prototype tool called BPMN2EVENTB.
Authors:Mr. Darshan Pandit Dr. Smitha Chowdary Mr.P.S.R. Patnaik Mr. Bhushan Shaharkar Mr.AkshayKumar Surde Abstract:Digitization keeps evolving hastily with enormous IoT devices. This Digitization era embraces many IoT applications from domains like smart ci-ties, smart energy, smart agriculture, smart healthcare, smart transportation & mobility, etc. These applications are much consumer oriented and are prone to frequently change according to the environmental change. Also, the different IoT devices from heterogeneous networks participate in the smart application which styles the IoT environment to be more complex and increases the number of interactions between devices and the application. The IoT device requires a novel software engineering approach with a set of rules & regulations to deal with continuous evaluation and decision making while moving through an IoT environment. This paper provides an insight into the agile model that boosts the support for consistent connectivity with a heterogeneous devices and flawless flow of its activities in IoT environments. It helps developer teams to work col-laboratively and organize the activities efficiently for achieving frequent deli-very to the stakeholders & end users.
Authors:Talha Naqash M. Hassan Tanveer Sajjad Hussain Shah Muhammad Salman Abstract:Software-Defined Network (SDN) consists of two layers; control and data layer that makes SDN more flexible and scalable. Open Flow protocol used for SDN, which makes it simpler and easy to optimize. In this paper, we developed a SABIDS for the Python-based controller (RYU) which detects the incoming traffic by taking their flow statistics, detects the malware flow statistics (by using the pattern match technique), and identifies the malicious flow. Also, it identifies the source IP of the incoming malicious traffic, and that specific IP can be blocked easily using the blacklist technique. This scheme enables the SDN Controller to learn about malicious traffic and avoid the potential losses like system failure or risk of being an attack.
Authors: Aditya Veer Mohit Gurav Shreyansh Dange Shubham Chandgude, Vaishali Wangikar Abstract: In the field of Knowledge Discovery in Databases (KDD) the effectiveness of association rules is important. Association Rules is a technique of data mining, wherein we identify the relationship between one item to another. For mining, the association rules Apriori algorithm is widely used. The idea of the Apriori algorithm is to find the frequent sets from a transactional database. Through the frequent sets, association rules are obtained, these rules must satisfy the minimum confidence threshold. This paper presents an improved method for deciding an optimum minimum support threshold and minimum confidence threshold, pruning of rules based on a contingency table, and finally the decision about whether to go for lift or confidence to get rid of uninteresting, misleading, and confusing association rules.
Authors:Fredrick Ishengoma, Leonard Mselle, Hector Mongi Abstract:The rising number of cell phone subscribers in Tanzania provides the government with a new platform for the provision of information and government services to people (thus m-Government). In Tanzania, the use of m-Government services is in the initial stages, and variables that affect its adoption are not yet understood. The goal of this research is to study the Critical Success Factors (CSFs) affecting the behavioural intention of citizens (BI) to adopt m-Govern-ment services in Tanzania. The study used the Mobile Services Acceptance Model (MSAM) and extended it to include external variables in the context of Tanzania. To collect primary data from users of m-Government services in Dar es Salaam and Dodoma towns, a survey questionnaire was used, and 253 re-sponses were collected. IBM-SPSS AMOS 23.0 program analyzed the data using Structural Equation Modeling (SEM). The findings of the study indicate that per-ceived usefulness, trust, perceived mobility, power distance, quality of service, awareness, perceived cost, personal initiatives and characteristics significantly influence the BI to adopt m-Government services. Perceived ease of use, was found to be statistically insignificant in predicting BI. Furthermore, the interplay between CSFs, discussion on theoretical and practical implications that follow from the results are presented.
Authors:Juan Moreno-Motta, Felipe Moreno-Vera, Frank A. Moreno Abstract:Currently, IoT has evolved to such an extent to extend to all corners of each place through devices that are connected to a network and generate information. In most cases, to be processed for a speci c purpose or storage as historical data; an IoT ecosystem is implemented to manage those tasks between di erent devices, frameworks, or applications. Besides, more complex IoT ecosystems require more complex architecture to manage the information ow, at this level, we found a problem called interoperability. This problem is not limited to the compatibility of adding/removing devices to an ecosystem, it is also expected that the information generated by devices and processed comply with a standard optimizing the data transmission. In this work, we present a new software architec ture pattern to avoid the problem of interoperability through the process of exchange information between devices and prevent store heterogeneous information coming from di erent layers of an IoT ecosystem.
Authors:Cheng Shen Wanli Xue Abstract:While Internet of Things (IoT) can bene t from machine learning by outsourcing model training on the cloud, user data exposure to an untrusted cloud service provider can pose threat to user privacy. Recently, federated learning is proposed as an approach for privacy preserving machine learning (PPML) for the IoT, while its practicability remains unclear. This work presents the evaluation on e ciency and privacy performance of a readily available federated learning framework based on PySyft, a Python library for distributed deep learning. It is observed that training speed of the framework is signi cantly slower that of the centralized approach due to communication overhead. Meanwhile, the framework bears some vulnerability to potential man-in-the-middle attacks at network level. The report serves as a starting point for PPML performance analysis and suggests the future direction for PPML frame work development.
Authors:Abhinav Reddy K. Sanjana Reddy D. Ranjani Yoganandam Palivela Srikar Sagar Basavaraju Soumya Menon Abstract:Forty percent of India's water supply is attributed to ground water. Muljipura, an agriculture-based village in the northern state of Madhya Pradesh, India, primarily utilizes groundwater and rainwater for irrigation, with a greater reliance on groundwater during non monsoon seasons. Due to over dependence on groundwater, there was a signi cant decline in groundwater levels between between 2007 and 2017. Now the rate of replenishment is unable to meet the demands of villagers. This pa- +E28per o ers a brief assessment of four appropriate technological solutions derived from literature and results obtained from Participatory Rural Appraisal and Human Centered Design methodological approaches. Both participatory-centric methods aided in collecting primary data to iden- tify salient characteristics of Muljipura village and speci c challenges faced by villagers. Primary data was combined with secondary data to assess four sustainable solutions. This study recommends an Internet of Things (IoT)-based appropriate technological solution to address rapid depletion of groundwater in the village of Muljipura.
Authors:John Colaco Rajesh lohani Abstract:There is a pressing need to be careful and ceaselessly screen the ascent in seawater level as people living along the Indian coastline are at risk of moving washed away because of the quick increasing speed of ascending in ocean level as it will prompt an enormous flood. This rise in seawater level is due to Global Warming and Climate Change. Measuring and monitoring the level of seawater through remote sensing is important to save humanity. Hence, remote sensing communication system satellite plays a major role in providing vital information about the rise in the seawater level. To get this vital information, a multi-signal active patch sensor array has to be incorporated on satellite for accurate and reliable sensing of and monitoring of data from the sea. This is achieved by designing a multi-band MIMO microstrip patch antenna sensor array in the frequency band of 1 GHz to 10 GHz resonating at 2.4 GHz, 6.8 GHz and 8.8 GHz for detection and monitoring rising sea water level. This work aim is to detect the hydrostatic pressure level of seawater in real-time through satellite communication.
Authors:K. Sridevi Dr. M A Saifulla Abstract:The control plane is the heart of the Software De ned Net- works, so it is mandatory that the work should go e ciently to increase the performance of the network. The challenge in the distributed control plane using multiple controllers is the even distribution of load among all the controllers. If the total load is distributed equally, the controller response time can be decreased and can achieve maximum throughput. This paper presents a Load Management Algorithm that chooses the required amount of load to shift from the heavy load controller to the light load controllers and allows parallel switch migrations in one cycle rather than multiple cycles. Experimental setup using RYU show that our approach can decrease the load of heavy load controller and better balance the control plane.
