Research Papers - Dept of Information Technology
Permanent URI for this collectionhttps://rda.sliit.lk/handle/123456789/593
Browse
15 results
Search Results
Publication Embargo Analyzing Payment Behaviors And Introducing An Optimal Credit Limit(IEEE, 2019-12-05) Bandara, H. M. M. T; Samarasinghe, D. P; Manchanayake, S. M. A. M; Perera, L. P. J; Kumaradasa, K. C; Pemadasa, N; Samarasinghe, PIdentifying an optimal credit limit plays a vital role in telecommunication industry as the credit limit given to customers is influence on the market, revenue stabilization and customer retention. Most of the time service providers offer a fixed credit limit for customers which may cause customer dissatisfaction and loss of potential revenue. Therefore, it is essential to determine an optimal credit limit that maintains customer satisfaction while stabilizing the company revenue. Clustering algorithms were used to group customers with similar payment and usage behaviors. Then the optimal credit limit derived for each cluster is applicable to all the customers within the cluster. In order to identify the most suitable clustering algorithm, cluster validation statistics namely, Silhouette and Dunn indexes were used in this research. Based on the scores generated from these statistics KMeans algorithm was chosen. Furthermore, the quality of the KMeans clustering was evaluated using Silhouette score and the Elbow method. The optimal number of clusters are identified by those validation statistics. The significance of this approach is that the optimal credit limits generated by these clustering models suit dynamic behaviors of the customer which in turn increases customer satisfaction while contributing to reducing customer churn and potential loss of revenue.Publication Embargo EyeDriver: Intelligent Driver Assistance System(IEEE, 2019-12-18) Gayadeeptha, P; Baddewithana, T. P; Pannegama, K. V; Samarakkody, C. S; Samarasinghe, P; Siriwardana, S“EyeDriver” is a driver assistance system that analyzes and provides real-time driver assistant data from four separate components. These main components are drowsiness detection and head pose estimation, over-speed detection, lane departure, and front collision avoidance. It is a compact product that included a Raspberry pi board, a USB camera module, Pi camera, and a TFT LCD. Since the “EyeDriver” is a first affordable aftermarket solution in Sri Lanka, it can be mounted and configured in any vehicle without any professional knowledge in less effort. Drowsiness detection and head pose estimation component will monitor the driver's eyes and keep track of whether the driver's head's position is inconsistent or deviated from the optimal position. In accordance with the road's recommended speed, the vehicle's actual speed is analyzed and if it is more than the permitted, the system makes a notification. It is done by the over-speed detection component. Lane departure component consists of assisting in keeping the vehicle stable on the desired lane on the road. Also, when the driver makes an intended lane change, the system provides a notification. The Front collision avoidance part will detect the frontal obstacle on the road and provide pre-collision/proximity warning notification. The notification makes according to the vehicle speed and distance between the object and the vehicles. The whole system is based on the Raspberry Pi 3 Model B+ board and the implementation of the system has been done by using OpenCV and Python.Publication Embargo Analysis and performance of CMA blind deconvolution for image restoration(Wiley Online Library, 2015-09) Samarasinghe, P; Kennedy, R. AIn this paper we study the applicability of classical blind deconvolution methods such as constant modulus algorithm (CMA) for blind adaptive image restoration. The requirements such as the source to be white, uniformly distributed and zero mean, which yield satisfactory convergence in the data communication application context, are revisited in the image restoration context, where a linear deblur kernel needs to be blindly adapted to compensate for an unknown image blur kernel with the objective to recover a source ground truth image. Through analysis and performance studies, we show that the performance of CMA is adversely affected by the intrinsic spatial correlation of natural images and by any deviation of their distribution from being platykurtic. We also show that decorrelation techniques designed to overcome spatial correlation cannot be effectively applied to rectify CMA performance for blind adaptive image restorationPublication Embargo Sri lanka driving license forgery detection(IEEE, 2017-12-21) Samarasinghe, P; Lakmal, L. K. P; Weilkala, A. V; Wickramarachchi, W. A. N. P. C; Niroshana, E. R. SAs there has been a significant increase in the number of identity thefts, governments and organizations have taken the detection of individual identities as a serious task. Sri Lanka has also experienced an increasing level of this issue in the recent past, especially in forging driving licenses. Since Sri Lanka Driving License (SLDL) has unique features, detection of fake driving licenses adopted by other countries cannot be directly applied for SLDL. As a developing country Sri Lanka cannot afford high cost scanning devices across the country to detect fake licenses. Overcoming these issues and addressing the requirement of counterfeit SLDL identification, in this research we came up with a cost effective, automatic and efficient image processing based SLDL identification system. Considering the unique features of SLDL, this system is able to identify counterfeit SLDL for all the types of driving licenses currently in use. Further, in this research novel image processing techniques have been applied in order to yield the highest accuracy through the system.Publication Embargo Deep learning based flood prediction and relief optimization(IEEE, 2019-12-05) Pathirana, D; Chandrasiri, L; Jayasekara, D; Dilmi, V; Samarasinghe, P; Pemadasa, NFlood is a major natural disaster that occurs recurrently in Sri Lanka. It is important to stay on alert and get early preparations to avoid unnecessary risks that cause damage to both life and property. This project developed a flood assistance application “DHARA” to support early flood preparation and flood recovery process. DHARA mobile application facilitates river water level prediction, safest evacuation route suggestion and provides relevant warnings and alert notifications and the web application provides affected area detection, victim and relief estimation to assist flood recovery management. This system is developed as a mobile application and a web application. A recurrent neural network architecture named Long Short Term Memory (LSTM), Convolutional Neural Network (CNN), a path finding algorithm namely A star (A*) algorithm and a clustering technique named Fuzzy Clustering are used for the development of the system. The system is verified with sample data related to “Wellampitiya” and “Kaduwela” area based on river “Kelanl”. The river water level prediction model has successfully predicted the water level 4 hours in advance. The verification results of the river water level prediction showed a satisfactory agreement between the predicted and real records with 85.4% accuracy.Publication Embargo Potential upselling customer prediction through user behavior analysis based on CDR data(IEEE, 2019-12-18) Manchanayake, S. M. A. M; Samarasinghe, D. P; Perera, L. P. J; Bandara, H. M. M. T; Kumaradasa, K. C; Premadasa, N; Samarasinghe, PUpselling is a valuable technique for increasing the profit margin of any service providing business domain. It plays a vital role in growth of a company. Among those companies, telecommunication industry is a prominent industry where upselling is highly influenced on churn reduction and stabilizing the customer base. As this increases the satisfaction of customers through adding products and services it is very effective in a marketing perspective. In a typical 4G LTE package marketing, customer will be offered to select a fixed package out of a set of pre-defined packages. The decision of selecting a suitable package by customer will be mainly an instinct driven decision due to lack of previous experience of using an LTE package. This will result in selecting an unsuitable package, which will lead to over usage of data. As a result, it will cause customer dissatisfaction and loss of potential income for the company. Therefore, it is essential for a telecommunication company to identify the customers who have the potential to upgrade their packages based on customer usage. In this research, potential package upgrades are predicted for LTE broadband users through a supervised learning method using different classification models. One of the key factors on obtaining a high accuracy model for classifying customers is to address the class imbalance problem that is present in the telecommunication data. The ratio between potential package upgrades and normal customers is highly skewed towards normal customers. This is addressed using SMOTE which is an oversampling method that creates synthetic samples using existing data points. Potential customers identified by a classification model trained by a dataset consist of usage behavior. The prediction results will be published to a dashboard that can be consumed by decision makers.Publication Embargo Screening Tool for Autistic Children(IEEE, 2019-01-23) Tittagalla, V. Y; Wickramarachchi, R. R. P; Chandrarathne, G. W. C. N; Nanayakkara, N. M. D. M. B; Samarasinghe, P; Rathnayake, P; Pemadasa, M. G. N. MAutism is a neurological disability that has been caused due to brain abnormality in a person. A person with Autism Spectrum Disorder(ASD) usually has difficulty in social and communication skills. In the past few years there hasn't been a proper way of identifying Autistic children in Sri Lanka. In this research paper, we will discuss how to identify an autistic child by considering mobile application with the following factors. Identify the eye contact, responsiveness to stimulus, analysis of vocal behavioral patterns and questionnaire. The above four factors will be the main key areas in screening process. This tool is created especially for identifying children with autism in rural areas in Sri Lanka. The major three areas eye contact, vocal behavior and responsiveness are the screening process is developed for proof of concept in this research.Publication Embargo “SenseA”-Autism Early Signs and Pre-Aggressive Detector Through Image Processing(IEEE, 2017-12-04) Gamaethige, C; Gunathilake, U; Jayasena, D; Manike, H; Samarasinghe, P; Yatanwala, TThis paper presents an efficient solution for the current problem of identifying early signs of autism and detecting pre aggressive behaviours using videos which can be used to produce a more convenient environment for autistic children and their caregivers. Early detection of autism spectrum disorder and its consequences play major role in intervention. Yet it often remains unrecognized and diagnosed in non-clinical environments because of unawareness and the lack of screening tools specific to the autism. At times, autistic children express their feelings through aggressive behaviour towards themselves or other children due to number of reasons such as failing to understand their own feelings, misunderstanding and severe distress. While an autistic child engages in physical aggression, an immediate response is required because the sibling or peer will likely react to the child's aggressive behaviour. There's no systematic approach to identify the pre-aggressive behaviour of autistic children in software engineering perspective. The proposed Autism Spectrum Disorder (ASD) early signs and pre-aggressive detection system is a software which provides automated solution to aforementioned problems using computer vision and machine learning libraries. It provides a level detection on early signs of autism by analyzing facial features and behaviour patterns in non-clinical perspective. Finally it detects the pre aggressive behaviours of autistic children and alerts the relevant authorized individual using a mobile notification.Publication Embargo Machine learning based automated speech dialog analysis of autistic children(IEEE, 2019-10-24) Wijesinghe, A; Samarasinghe, P; Seneviratne, S; Yogarajah, P; Pulasinghe, KChildren with autism spectrum disorder (ASD) have altered behaviors in communication, social interaction, and activity, out of which communication has been the most prominent disorder among many. Despite the recent technological advances, limited attention has been given to screening and diagnosing ASD by identifying the speech deficiencies (SD) of autistic children at early stages. This research focuses on bridging the gap in ASD screening by developing an automated system to distinguish autistic traits through speech analysis. Data was collected from 40 participants for the initial analysis and recordings were obtained from 17 participants. We considered a three-stage processing system; first stage utilizes thresholding for silence detection and Vocal Activity Detection for vocal isolation, second stage adopts machine learning technique neural network with frequency domain representations in developing a reliant utterance classifier for the isolated vocals and stage three also adopts machine learning technique neural network in recognizing autistic traits in speech patterns of the classified utterances. The results are promising in identifying SD of autistic children with the utterance classifier having 78% accuracy and pattern recognition 72% accuracy.Publication Embargo Self-speech evaluation with speech recognition and gesture analysis(IEEE, 2018-10-02) Shangavi, S; Jeyamaalmarukan, S; Jathevan, A; Umatharsini, M; Samarasinghe, PSpeaking helps people to improve their communication, public speaking and leadership skills. There are two main techniques that help a speaker to deliver a meaningful speech. The techniques are voice transition which expresses a verbal message and gestures that convey the message to an audience. A famous organization to help and improvise speech is Toastmasters. Their systems of evaluation are such as, Tracking Filler words, Usage of Redundant words and Phrases, Checking Grammar and Pronunciation, Usage of Body Movements and Gestures, Tracking Vocal Variations and Time Management. If an ordinary person wants to self-evaluate his or her speech, that person has to be a member of a Toastmasters Club or any other speech improvising organization. By using our application, it is possible for a person to evaluate his or her own speech without depending on an organization. All the above-mentioned criteria in manual evaluation processes are included in this application. Since nowadays mobile applications are frequent in use, our system is proposed in Android Platform. Several techniques and methods are used to interconnect with Android such as OpenCV, Microsoft Cognitive Services and MATLAB in order to achieve the objectives of the application. Acoustic Model, Support Vector Model (SVM), Hidden Markov Model (HMM) are some models used to build the application more efficient by giving approximately accurate results.