Asian Journal of Convergence in Technology
Not a member yet
868 research outputs found
Sort by
Handwritten Character Recognition using Convolution Neural Networks in Python with Keras
Get PDFIn the field of Deep Learning for Computer Vision, scientists have made many enhancements that helped a lot in the development of millions of smart devices. On the other hand, scientists brought a revolutionary change in the field of image processing and one of the biggest challenges in it is to identify documents in both printed as well as hand-written formats. One of the most widely used techniques for the validity of these types of documents is ‘Character Recognition’. This project seeks to classify an individual handwritten word so that handwritten text can be translated to a digital form. It demonstrates the use of neural networks for developing a system that can recognize handwritten English alphabets. In this system, each English alphabet is represented by binary values that are used as input to a simple feature extraction system, whose output is fed to our neural network system. The CNN approach is used to accomplish this task: classifying words directly and character segmentation. For the former, Convolutional Neural Network (CNN) is used with various architectures to train a model that can accurately classify words. For the latter, Long Short Term Memory networks are used with convolution to construct bounding boxes for each character. We then pass the segmented characters to a CNN for classification, and then reconstruct each word according to the results of classification and segmentation. 
Review on Phishing Attack Detection Techniques
Get PDFPhishing attacks capitalize on human errors and target the vulnerabilities formed due to it. Most of the attacks are aimed at stealing private information from users, which spread via different mechanisms. There is no single solution to this problem to effectively nullify all the attacks but multiple techniques have been developed to defend against these attacks. This paper reviews the work on the detection of phishing attacks. In this paper, we aim to study the techniques which mainly detect and help in preventing phishing attacks rather than mitigating them. A general run-through of the most successful techniques for phishing attack detection has been presented here
Impact of Social Media on Big Data and Green Computing and Solution
Get PDFBig data‘s one of the major challenge is volume of data. With high speed 4G internet and with the occurrence of smart hand held devices like Smart cell phones, iPads, Notebook computers, increased the number of internet users, that too online social user. This has led to online social data generated in tremendous rate never seen before. This huge data generated encountered various challenges like storage, maintenance and green computing for software industries. So, in this paper I proposed an algorithm which provides a form of optimal solution to control unwanted data generation from social media interaction and also discussed its impact on storage and on Green Computing and respective solution
Using Machine Learning, Image Processing & Neural Networks to Sense Bullying in K-12 Schools
Get PDFWe all have heard about bullying and we know that it is an immense challenge that schools have to tackle. Many lives have been ruined due to bullying and the fear it implants into students' mind has caused many of them to go into depression which can lead to suicide. Traditional methods [1] need to be accompanied with modern technology to make the method more effective and efficient. If real time alerts are to school staff, they can identify the perpetuator and extricate the victim swiftly. It this proposed method an AI based solution is implemented to monitor students using standard school surveillance technologies and CCTV to maintain a decorum and safe environment in the school premise. Also the proposed method utilizes other unstructured sources such as attendance records, social media activity and general nature of the students to deliver quick response. Artificial Intelligence (AI) techniques like Convolutional Neural Networks (CNN), which includes image processing capabilities, logistic regression methods, LSTM (Long short-term memory), and pre-trained model Darknet-19 is used for classification. Further, the model also included sentiment analysis to identify commonly used abuse terms and noisy labels to improve overall model accuracy. The model has been trained and validated with the realistic data from all the sources mentioned and has achieved the classification accuracy of 87% for detecting any sign of bullying
Significance-Driven Logic Compression for Energy Efficient Multiplier Design
Get PDFApproximate Arithmetic is a new design paradigm that is being used in many applications which are tolerant to imprecision and do not require accurate results. It can reduce circuit complexity, delay and energy consumption by relaxing accuracy requirements. The partial product bit matrix can be reduced based on their progressive bit significance using a Significance Driven Logic Compression(SDLC) approach. Further, the complexity of the approximate multiplier can be reduced by using Approximate adders in place of exact adders in the accumulation method. Removing some of the transistors from an accurate adder will result in an approximate adder. By using approximate adders which have less number of transistors, the power, propagation delay, and the switching capacitance can be reduced. In this paper, approximate multipliers are implemented using different approximate adders and they are compared with an exact multiplier in terms of power, delay and energy savings
Search Engine Marketing Using Search Engine Optimisation.
Get PDFToday, E-commerce is booming, so is the number ofsales on online market as compared to the offline market which isalmost constant and also saturated. There are two types ofwebsites in today’s market, first, the genuine websites trying tomake their mark in market and second, company websites withgood brand name who already have a head start because of theirbrand value. These websites are always competing against eachother in the market, but the website with a large customer base isobviously the one which is making large profits. So this narrowsdown the whole discussion to one point, customers! When aregular customer with no prejudice about the online websitesmakes a search on the search engine that is where it all starts, theresults showing numerous of websites but in a specific order andthe customer tends to select his website from the first few options.This is what we are going to discuss in the paper that how doesyour website get in those first few results. Also, this huge increasein the number of websites has led to creation of a new industry,‘Search Engine Marketing’ about which we will be discussing inthe paper. Now, as of today, the goal is to increase websiteranking to display the website in top search results
Smart antenna array processing techniques and implementation of Spatio-temporal sampling and Equality check algorithm
Get PDFResearch and development in smart or adaptiveantenna array is always in progress to estimate direction ofarrival (DOA) accurately and to form the beam in the directionof DOA. This paper presents different temporal and spatial arrayprocessing techniques used for DOA estimation. Temporal classcontains all algorithms which use temporal properties of thedesired signal during the adaptation process. This can be atraining sequence as well as the constant envelope property of thesignal to be received. The antenna weight vectors are directlydetermined from the received signals at the antenna elements andthe temporal signal properties, where no channel estimation isperformed. Temporal processing algorithms include LMS, RLS,SQRLS, DMI, SMI, CMA algorithms. Spatial algorithms useknowledge about the properties of the array manifold todetermine the DOAs of the incident signals. Afterwards theantenna weight vectors are determined and the signals areseparated and further processed by any detector. Spatialprocessing algorithms include MUSIC, ESPRIT, UNITARYESPRIT, SAGE algorithms.In this paper by studying the above method onedifferent method i.e. ‘spatio-temporal sampling and equalitycheck’ method is developed and implemented using MATLABSIMULINK and the performance is checked for differentcombination of input signal and noise signal
SMART SHOPPING BASKET
Get PDFIn modern India where people talk about smartcities and India as a digitizing nation, it is very important thatwe start with necessities of citizens. Grocery shopping is onesuch necessity that requires modernization. There has alwaysbeen a long queue at the billing counters in shopping malls.This activity many times consumes lot of time which results infrustration amongst customers. This problem is faced byeveryone. Especially in India where there is a lot of populationand not enough billing counters. There are online grocerystores available, but they have their own disadvantages likeminimum bill value etc. To overcome this problem an idea hasbeen developed which can be implemented in all the shoppingmalls to save customer’s precious time and simplify the billingprocess.In a country like India where cost is an important parameter, itis important to keep in mind the manufacturing cost of theproduct. If the cost is low without compromising thefunctioning, the model can be implemented on a larger scale.Considering this the model has been built with the mostcommonly available and efficient hardware. This paperdescribes the idea of smart shopping basket and how it can bebuilt using basic cost-efficient hardware. In the proposedmodel, Radio Frequency Identification (RFID) has been usedto detect grocery items. Every product will have a RFID tagand every cart will have a reader. A RFID reader is used todecode the tag and display the information on a small LiquidCrystal Display (LCD). Technical working of the circuit withhardware specifications has been discussed and ways ofoptimization have been suggested. Different variants of thisidea have been studied from research papers and a simplifiedand effective model with working has been proposed in thispaper. 
ENHANCED PHYSICALLY CHALLENGED SCOOTERS WITH AUTOMATIC BALANCING WHEEL ADJUSTMENT
Get PDFThis paper reports on enhanced physicallychallenged scooters with automatic balancing wheeladjustment. It mainly aims to focus on retrofitting specialfeatures in the existing scooters for physically challengedpeople. By automating the adjustment of balancing wheelattached parallel to the rear running wheel, the active lifecycle of the vehicle can be increased. It is intended to reducethe wear and tear losses of the balancing wheels and torestorethe mileage of the vehicle
Social Media Text Mining for Decision Support in Natural Disaster Management in Sri Lanka
Get PDFWith the popularity of the internet andsmart devices, social media is viral today among individuals inalmost all the ages which help them to create and share theirpersonal feelings, experiences, ideas, as well as informationwith others connected to them over a computer, mediatedtechnology. Due to this nature when there are emergencies andnatural disasters these social media applications tend to beflooded with content generated from the public who affected,who are looking for their family members and friends, who arelooking for information as well as with the people engage inhumanitarian activities.Therefore, social media has become thefirst to generate related information when there is acatastrophic event before any of news sites or governmentbodies engage in disaster management. This social mediacontent is quick accurate and subjective during disastersituations, therefore,can be used as an asset to reduce risk andbuild awareness among the public about the disaster as well asto provide decision making support to relief efforts. Thisresearch focuses on building decision making support usingsocial media content generated during disaster situations in theSri Lankan context. Mainly the content will be tweets posted bythe public during a natural disaster and consisting of textwritten in English. Therefore,situational awareness buildingwill be done using text mining techniques in this study since thecontent is unstructured