Research Papers - Dept of Information Technology
Permanent URI for this collectionhttps://rda.sliit.lk/handle/123456789/593
Browse
12 results
Filters
Advanced Search
Filter by
Settings
Search Results
Publication Embargo Sinhala Part of Speech Tagger using Deep Learning Techniques(IEEE, 2022-12-21) Sathsarani, M.W.A.R.; Thalawaththa, T.P.A.B.; Galappaththi, N.K.; Danthanarayana, J.N.; Gamage, ANatural Language Processing (NLP) is a sub-field of Artificial Intelligence (AI) that consists of a collection of computational methods motivated by theory for the automated classification and reflection of human languages. The foundation for many sophisticated applications of NLP, including named entity recognition, sentiment analysis, machine translation, in-formation retrieval, and information processing, is laid by Part of Speech (POS) tagging, which is part of the lexical layer of NLP systems. In contrast to English, French, German, and other languages from the same geographical region, the development of high-accuracy, stable POS taggers for the Sinhala language is still in its early stages. Hence, Sinhala is identified as a low-resource language. The main objective of this research is to create a POS tagger for the Sinhala language to solve this issue. An innovative and novel strategy that has never been used with the Sinhala language has been designed. This approach has been suggested specifically to evaluate the possibility of enhancing the accuracy compared to other methodologies. So, deep learning algorithms have been applied in this study, which has a significant impact on improving tagger performance. First, highly accurate individual classifiers for primary POS tags were implemented, and then they were combined into one composite model. As expected, all individual classifiers and the final composite model have achieved a higher accuracy level. Thus, it demonstrates that the proposed solution using deep learning algorithms outperformed other methods, such as rule-based and stochastic, in terms of accuracy.Publication Embargo English Language Trainer for Non-Native Speakers using Audio Signal Processing, Reinforcement Learning, and Deep Learning(IEEE, 2021-12-02) Jeewantha, H. C. R.; Gajasinghe, A. N; Rajapaksha, T. N; Naidabadu, N. I; Kasthurirathna, D.; Karunasena, A.Lack of basic proficiency and confidence in writing and speaking in English is one of the major social problems faced by most non-native English speakers. Although the general adult literacy rate in Sri Lanka is above average by world standards, the English literacy rate is just 22% among the Sri Lankan adult population. Many individuals face setbacks in achieving their career and academic goals due to these language barriers. In a world where English has become a compulsory requirement to pursue higher education, career development, and performing day-to-day activities, "English Buddy" is a software solution developed to enhance the English learning experience of individuals in a more personalized and innovative way. The system provides an all-in-one solution while filling major research and market gaps in existing solutions in the e-learning domain. The system consists of a personalized learning environment, automated pronunciation error detection system, automated essay evaluation process, automated descriptive answer evaluation component based on semantic similarity, and real-time course content rating system. English Buddy is implemented using state-of-the-art technologies such as Audio Signal Processing, Reinforcement Learning, Deep Learning, and NLP. The LSTM, Sentiment Analysis, and Siamese network models have gained accuracy scores of 0.93, 0.92, and 0.81 respectively. Further, the UAT results proved that the personalized recommendations and pronunciation error detection processes are accurate and reliable. This research has been able to overcome the limitations of most existing solutions that follow traditional approaches and provide better results compared to the recent studies in the e-learning research domain.Publication Embargo Deep Learning Approach for Designing and Development of Risk Level Indicator for Patients with Lung Diseases(IEEE, 2022-02-23) Chathurika, K. B. A. B; Gamage, A"Lung disease" as a medical term, discusses as several disorders that affects both lungs. There are different types of lung disease like Asthma, lungs infections like Influenza, Pneumonia, Tuberculosis, and numerous other types of breathing problems including Lung cancers. These lung diseases can be the main reason for failure in breathing. Due to COVID19 pandemic, Pneumonia and COVID19 were highlighted mostly as fatal diseases if not detected on time. Newly identified COVID19 diseases has caused many deaths and confirmed detections reported worldwide, followed with a greatest risk to community wellbeing, especially for patients with lung diseases. Process of developing a clinically accepted vaccine or specific therapeutic drug for this disease are not finalized, which will contribute to the expansion of actual prevention action plans. Thus, methods to detect lung illness accurately and efficiently is important. Proposed solution will easily and precisely detect the risk level of patients with these two lung diseases Pneumonia and COVID19 using a mobile application with chest radiography (Chest X-rays), which is considered as a cheap, easy to access and speedy manner. Proposed solution will identify, classify and evaluate the risk level of the patient suffering with the use of Image Processing, Machine Learning techniques and Convolutional Neural Networks. So, anybody who use the proposed solution may have the ability to have a precious decision about own medical condition accurately, quickly with low cost. Proposed solution can calculate severity level of a patient with more than 97% accuracy with chest radiography analysis together with patient’s current symptoms and breath holding time evaluation.Publication Embargo Coconut Disease Prediction System Using Image Processing and Deep Learning Techniques(Institute of Electrical and Electronics Engineers Inc., 2020-12-09) Nesarajan, D; Kunalan, L; Logeswaran, M; Kasthuriarachchi, S; Lungalage, DCoconut production is the most important and one of the main sources of income in the Sri Lankan economy. The recent time it has been observed that most of the coconut trees are affected by the diseases which gradually reduces the strength and production of coconut. Most of the tree leaves are affected by pest diseases and nutrient deficiency. Our main intensive is to enhance the livelihood of coconut leaves and identify the diseases at the early stage so that farmers get more benefits from coconut production. This paper proposes the detection of pest attack and nutrient deficiency in the coconut leaves and analysis of the diseases. Coconut leaves monitorization has been taken place after the use of pesticides and fertilizer with the help of the finest machine learning and image processing techniques. Rather than human experts, automatic recognition will be beneficial and the fastest approach to identify the diseases in the coconut leaves very efficiently. Thus, in this project, we developed an android mobile application to identify the pests by their food behaviors, pest diseases and the nutrition deficiencies in the coconut trees. As an initial step, all datasets for image processing technology met pre-processing steps such as converting RGB to greyscale, filtering, resizing, horizontal flip and vertical flip. After completing the above steps, the classification was performed by analyzing several algorithms in the literature review. SVM and CNN were chosen as the best and appropriate classifier with 93.54% and 93.72% of accuracy respectively. The outcome of this project will help the farmers to increase the coconut production and undoubtedly will make a revolution in the agriculture sector.Publication Open Access Gesture driven smart home solution for bedridden people(Association for Computing Machinery, 2020-09-21) Jayaweera, N; Gamage, B; Samaraweera, M; Liyanage, S; Lokuliyana, S; Kuruppu, TConversion of ordinary houses into smart homes has been a rising trend for past years. Smart house development is based on the enhancement of the quality of the daily activities of normal people. But many smart homes have not been designed in a way that is user friendly for differently-abled people such as immobile, bedridden (disabled people with at least one hand movable). Due to negligence and forgetfulness, there are cases where the electrical devices are left switched on, regardless of any necessity. It is one of the most occurred examples of domestic energy wastage. To overcome those challenges, this research represents the improved smart home design: MobiGO that uses cameras to capture gestures, smart sockets to deliver gesture-driven outputs to home appliances, etc. The camera captures the gestures done by the user and the system processes those images through advanced gesture recognition and image processing technologies. The commands relevant to the gesture are sent to the specific appliance through a specific IoT device attached to them. The basic literature survey content, which contains technical words, is analyzed using Deep Learning, Convolutional Neural Network (CNN), Image Processing, Gesture recognition, smart homes, IoT. Finally, the authors conclude that the MobiGO solution proposes a smart home system that is safer and easier for people with disabilitiesPublication 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 Facial emotion prediction through action units and deep learning(IEEE, 2020-12-10) Nadeeshani, M; Jayaweera, A; Samarasinghe, PWith the recent advancements in deep learning techniques, attention has been given to training and testing facial emotions through highly complex deep learning systems. In this paper we apply machine learning techniques which require less resources to produce comparable results for emotion prediction. As the underlying technique for the emotion prediction in this research is based on clinically recognized Facial Action Coding System (FACS), a further analysis is given on the contribution of each of the Action Units (AUs) for the predicted emotion. This analysis would complement, strengthen and be a main resource for addressing many different health issues related to facial muscle movements.Publication Embargo Pubudu: Deep learning based screening and intervention of dyslexia, dysgraphia and dyscalculia(IEEE, 2019-12-18) Kariyawasam, R; Nadeeshani, M; Hamid, T; Subasinghe, I; Samarasinghe, P; Ratnayake, pDyslexia, Dysgraphia and Dyscalculia are significant learning disabilities that affect around 10% of children in the world. Despite the advancement of technology literacy in the community, limited attention has been given for screening and intervention of these disabilities using mobile applications in Sri Lanka. In this research, one of the first deep learning and machine learning based mobile applications, named “Pubudu” was developed for screening and intervention of dyslexia, dysgraphia and dyscalculia supporting local languages. In “Pubudu” we have followed up clinical screening and diagnostic procedures recommended by health professionals for screening and intervention. The screening of dyslexia, letter dysgraphia and numeric dysgraphia was carried out using deep neural network and the screening for dyscalculia was carried out using machine learning techniques. Intervention techniques are implemented using gamified environments. System testing was carried out using 50 differently abled children and 50 typical children. With the initial dataset 88%, 58%, 99% screening accuracies are achieved in neural networks for letter dysgraphia, dyslexia and numeric dysgraphia screening while dysgraphia, whereas 90% accuracy was achieved for dyscalculia. Handwritten letters and numbers were fed as inputs to CNN model in letter dysgraphia and numeric dysgraphia while embedded audio clips of letter pronunciation were fed in to voice recognition CNN model in dyslexia. “Pubudu” shows significant potential for screening and intervention of dyslexia, dysgraphia and dyscalculia in local languages motivating children and interactively making them able and would be an enabling app for most of the underprivileged children in Sri Lanka.Publication Embargo OMNISCIENT: A Branch Monitoring System for Large-scale Organizations(2020 2nd International Conference on Advancements in Computing (ICAC), SLIIT, 2020-12-10) Jayasekara, T.; Omalka, K.; Hewawelengoda, P.; Kanishka, C.; Samarasinghe, P.; Weerasinghe, L.Omniscient is a system that enables higher-level management of massive organizations to remotely monitor and scrutinize the activities that take place in the branches from the head office itself by providing exclusive insight in the form of detailed reports on the employees’ behaviour and performance daily, weekly and monthly. The system further monitors the branch and provides reports on any suspicious behaviour and also on the customers’ activity within the branch premises. Omniscient rates the customer’s level of satisfaction by capturing the customer’s facial expressions and analyzing their emotions while they are being served. The employee face and dress recognition models have accuracies of 90.90% and 87.00% respectively while, employee activity detection has an accuracy of 89.00%. Customer emotion and miscellaneous activities detection models have the accuracies of 91.50% and 83.00% respectively. All of the aforementioned procedures were made possible by systematically analyzing the IP camera video footage obtained throughout the day to analyze the work productivity and performance of the branch as accurately as possible using deep learning and modern visual computing techniques like CNN, OpenCV, Haar Cascade classifier, face recognition, Dlib and Darknet.Publication Embargo Gamified Smart Mirror o Leverage Autistic Education – Aliza(2020 2nd International Conference on Advancements in Computing (ICAC), SLIIT, 2020-12-10) Najeeb, R.S.; Uthayan, J.; Lojini, R.P.; Vishaliney, G.; Alosius, J.; Gamage, A.Autism is a neurodevelopmental disorder that causes difficulties in communication, emotional responsiveness and social skills. There has been a global increase rate in autism and lack of resources locally to educate ASD children. As this condition affects children at an early stage, it remains a challenge in learning. Even though today's world there are ample of teaching methods and technologies, people are unaware of the use and impact of them. This paper presents “Aliza” Gamified smart mirror to teach basic education for autism children. “Aliza” consists of four core components such as writing mentor for pre-writing, math tutor for mathematics, verbal trainer for speech and attentiveness tracker for emotion detection. These components assist and enhance their competency in education. The users of the “Aliza" will be constantly monitored and evaluated during their training using Convolutional Neural Network (CNN). The interactive games are given to impact their learning process while the generated report from the Deep Learning evaluation system can acquaint parents and the tutors with the progress of the children. Through this research, it is expected to improve autistic children's basic education with assistance of “Aliza".