Browsing by Author "Gamage, A"

Now showing 1 - 20 of 29

Open Access
Agro-Genius: Crop Prediction Using Machine Learning
(2019-10) Gamage, A; Kasthurirathna, D
This paper present a way to aid farmers focusing on profitable vegetable cultivation in Sri Lanka. As agriculture creates an economic future for developing countries, the demand of modern technologies in this sector is higher. Key technologies used for this problem are Deep Learning, Machine Learning and Visualization. As the product, an android mobile application is developed. In this application the users should input their location to start the prediction process. Data preprocessing is started when the location is received to the system. The collected dataset divided into 3 parts. 80 percent for training, 10 percent for testing and 10 percent for validation. After that the model is created using LSTM RNN for vegetable prediction and ARIMA for price prediction. Finally, for given location profitable crop and predicted future price of vegetables are shown in the application. Other than the prediction, optimizing for multiple crop sowing according to the user requirements and visualizing cultivation and production data on map and graphs are also given in the application. This paper elaborates the procedure of model development, model training and model testing.
Embargo
AI and Machine Learning Based E - Learning System For Secondary Education
(IEEE, 2022-07-18) Wijayawardena, G. C. S; Subasinghe, S. G. T. S; Bismi, K. H. P; Gamage, A
One of the key functions directly shifted to online platforms under COVID-19 is education. The paper is about an E-learning system for secondary education in Sri Lanka. Learners and teachers can access information, resources, and tools through an E-Learning system, which is a Learning Management System that integrates a number of online activities. The main functions provided through the proposed system are chatbot, final grade prediction and weak area prediction of the students. Chatbots are becoming increasingly popular in a wide range of applications, especially in those that provide intelligence support to the user, according to recent research. So, in order to speed up the aid process, these systems are often integrated with Chatbots, which can quickly and accurately read the user's questions. This paper describes the implementation of a Chatbot prototype in the educational domain: a system for providing support to students. In the beginning, the goal was to design a special architecture and communication model that would help students get the proper answers. The final grade prediction component plays major role in the system. Because when the students are graded by their marks, they can review which areas that they have to improve and work on them. This is helpful for students as well as teachers. Weak area prediction also plays a significant role, because it can help to find out the weak areas of each subject and generate Individual Student Progress Plans to predict the students’ weak subjects and the subject areas of the students. This motivates students to get higher marks easily because this part is mainly focused on weak areas of students and improve those weak areas by providing several learning activities. These are the major parts of this system to have a good E-learning system for both students and Teachers.
Embargo
Analyzing Fisheries Market, Shrimp Farming & Identifying Fish Species using Image Processing
(IEEE, 2022-12-09) Sumeera, S; Pesala, N; Thilani, M; Gamage, A; Bandara, P
The fisheries industry is vital to the Sri Lankan economy because it provides a living for more than 2.5 million coastal communities and meets more than half of the country’s animal protein needs. Today, the fishery community in Sri Lanka is facing several grant problems. Among them, not getting a decent fish price for their harvesting, the inability to identify diseases in shrimp cages in the early stages, and the inability to identify fish species by observing their external appearance. This research developed a prototype mobile application “Malu Malu” to avoid the above-mentioned problems. It facilitates to the prediction of market fish prices, identifying shrimp diseases in their early stages, and identifying fish species by observing their external appearance. The proposed predictive models of the “Malu Malu” contains three main models developed using inseption V3 Convolutional Neural Network (CNN) model for image classification and Linear Regression is used for creating a model for predictions. The experimental results of these models showed above 85% of accuracy.
Embargo
ARROW: A web-based employee turnover analysis tool for effective human resource management in large-scale organizations
(IEEE, 2017-09-14) Weeramanthrie, T. T; Thilakumara, C. N; Wijesiri, K. N. A. C; Fernando, N. I; Thelijjagoda, S; Gamage, A
To gain the competitive advantage, organizations need to adapt to the dynamic market. Therefore, many researchers have tried to find different ways for adapting to competitive conditions. Most of these research have finally ended up focusing on the human resource, which is the major and important resource in any organization. Currently human beings are treated as assets rather than resources. The System, ARROW is a unique web application developed to satisfy the requirements of company management in employee understanding process. The main objective of the system, ARROW is to fulfil the gap between employees' past, present and future behavior and the management's ability to understand the behavior of the organization's employees at the HR practices. Natural Language Processing and Data Mining techniques were used to accomplish the main objective.
Embargo
BlossomSnap: A Single Platform for all Anthurium Planters Based on The Sri Lankan Market
(Institute of Electrical and Electronics Engineers, 2022-10-15) Rathnayake, R.M.S.T; Tharika Pramodi, M.L.A.D.; Gayathree, I. R; Rashmika, L.K.R; Gamage, M; Gamage, A
The popular and extensively grown flowering plant known as the Anthurium is prized for its beauty. In Sri Lanka, anthuriums have a substantial international market. Although it is a significant field that can be further developed by expanding the market, but it has led to a lack of attention, resources, and a moderate cost of production, as well as from the absence of an appropriate market channel, all of which have led to lower productivity and quality. As a result, Anthurium growers have numerous challenges both in terms of production and marketing. This paper introduces a novel mobile application 'BlossomSnap' which involves automating and significantly enhancing the outdated manual process. Using natural language processing, machine learning, and deep learning approaches, the proposed system analyzes the diseases, pests, varieties, and the highest quality plants to create a more secure growing environment. It will provide high-quality, cost effective, and timely services. The first step of anthurium plant disease and pest diagnosis is carried out using image processing, deep learning, and machine learning technologies. In order to identify the infection stage, the following steps involve extracting, classifying, and detecting images of Anthurium flowers and leaves. The accuracy was checked by comparing actual results taken from experts with the predicted results obtained from the proposed system. 'BlossomSnap' achieves an average accuracy of more than 80% and produces a better overall result. An in-place chatbot technology is intended to assist new planters with their problems. The Anthurium plant variety and quality detection methodology is used in concert with to determine the optimum market opportunity.
Embargo
Carbon Emission Optimization Using Linear Programming
(IEEE, 2022-12-09) Magenthirarajah, V; Gamage, A; Chandrasiri, S
In this fast-growing modernization, excess carbon emission plays a crucial role in climate change. Targeting and experimenting with sustainable ways of Carbon neutrality and management is the pathway toward a greener society. Data show that factories and industries take a high market stake in carbon emission and management. In actions, Governments defined a limit for carbon emissions to each organization which is called carbon credit. Every organization must focus on reducing carbon emissions. This is a critical task for each organization, In some cases, it is still not possible to explore other sustainable options. An innovative solution proposed for the above scenario is to implement a real-time platform that can provide insights into the most up-to-date emission statistics of the organization. This paper provides advanced analytics and precise proactive planning and actions in the simplest form and a discussion on future elaborations and insights about conclusions. By finding the minimum optimal emission values of each emission source, organizations can maintain carbon emissions without exceeding their carbon credit. Also, how industries and factories can create a smart carbon optimization system that can create an even greener society.
Open Access
Combined Approach of Supervised and Unsupervised learning for Dog Face Recognition
(IEEE, 2021-04-02) Weerasekara, D. T; Gamage, A; Kulasooriya, K. S. A. F
One would be surprised to hear the lost dog rates around the world. Even though it is something that one doesn't ponder a lot about, lost dogs are a problem that most dog owners fear. Dogs provide humans with companionship, protection, and unconditional love, and to the dogs; their whole world revolves around their owner and their family members. Therefore, when a pet dog goes missing, not only the dog owner but also the pet dog is affected. Unfortunately, in Sri Lanka, a lost dog being found is a very rare occurrence. A reason for this can be pointed out as the lack of an easily-accessible, public platform for lost dogs. In this research project, a solution to this problem has been implemented using image processing. This research study is about image classification and recognition using the Convolutional Neural Network (CNN) or also known as Shift Invariant or Space Invariant Artificial Neural Network (SIANN) by using TensorFlow framework as well as Keras library. The VGG16 model was customized for being used feature extraction. The implementation was a combination of both Machine Learning and Deep Learning. The platform to upload the found dog is also a continuous and inter-related subcomponent that provides a happy and healthy life for stray dogs too. That idea is providing them a higher chance to find a safe place to survive and also a home where they will be loved. The results are discussed in terms of the accuracy of the image recognition and classification in percentage. Each group of dogs get around 90% accuracy or above.
Embargo
Computer Vision and NLP based Multimodal Ensemble Attentiveness Detection API for E-Learning
(IEEE, 2021-04-21) Wijeratne, M. D; Lakmal, R. H. G. A; Geethadhari, W. K. S; Athalage, M. A; Gamage, A; Kasthurirathna, D
Attention is the fundamental element of effective learning, memory, and interaction. Learning however, with the evolvement of technologies in the modern digital age, has surpassed traditional learning systems to more convenient online or e-learning systems. Nevertheless, unlike in the traditional learning systems, attention detection of a student in an e-learning environment remains one of the barely explored areas in Human Computer Interaction. This study proposes a multimodal ensemble solution to detect the level of attentiveness of a student in an e-learning environment, with the use of computer vision, natural language processing, and deep learning to overcome the barriers in identifying user attention in e-learning. The proposed multimodal captures, processes, and predicts user attentiveness levels of individual students, which are subsequently aggregated through an ensemble model to derive an overall outcome of better accuracy than individual model outcomes. The final outcome of the ensemble model produces a range of percentages, within which the attentiveness level of the student lies during a single online lesson. This range is consequently delivered to the users through an Application Programming Interface.
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.
Embargo
Design of a kinetic energy harvester for elephant mounted wireless sensor nodes of jumbonet
(IEEE, 2016-12-04) Wijesundara, M; Tapparello, C; Gamage, A; Gokulan, Y; Gittelson, L; Howard, T; Heinzelman, W
In areas where the habitats of elephants and humans are rapidly encroaching on each other, real-time monitoring of the elephants' locations has the potential to drastically improve the co-existence of elephants and humans, resulting in reduced deaths in both groups. However, as tagging (using GPS collars) elephants to obtain such location information is difficult and costly, it is important to ensure very long lifetimes of the tags, which can only be achieved using energy harvesting. In this paper, we present a kinetic energy harvester that uses magnetic levitation and ferro fluid bearings to generate energy from an elephant's movements. In order to determine the feasibility of using this kinetic energy harvester for powering the tags on elephants, we obtained real acceleration data collected from an Asian elephant over a 10 day period, and this data was then used to tune the system to maximize the harvested energy. Using experimentally validated analytical and simulation models, and the actual elephant acceleration data, we find that our prototype can generate 88.91J of energy per day. This energy is not only sufficient to power the tags to acquire and transmit locations 24 times a day to a distance of 114Km (line of sight), but provides a surplus of at least 35.40J, which can be used to increase the frequency of position updates or to support alternative communication options such as GPRS. Therefore, this shows the viability of long-term tracking of elephants.
Embargo
Design of a kinetic energy harvester for elephant mounted wireless sensor nodes of jumbonet
(IEEE, 2016-12-04) Wijesundara, M; Tapparello, C; Gamage, A; Gokulan, Y; Gittelson, L; Howard, T; Heinzelman, W
In areas where the habitats of elephants and humans are rapidly encroaching on each other, real-time monitoring of the elephants' locations has the potential to drastically improve the co-existence of elephants and humans, resulting in reduced deaths in both groups. However, as tagging (using GPS collars) elephants to obtain such location information is difficult and costly, it is important to ensure very long lifetimes of the tags, which can only be achieved using energy harvesting. In this paper, we present a kinetic energy harvester that uses magnetic levitation and ferro fluid bearings to generate energy from an elephant's movements. In order to determine the feasibility of using this kinetic energy harvester for powering the tags on elephants, we obtained real acceleration data collected from an Asian elephant over a 10 day period, and this data was then used to tune the system to maximize the harvested energy. Using experimentally validated analytical and simulation models, and the actual elephant acceleration data, we find that our prototype can generate 88.91J of energy per day. This energy is not only sufficient to power the tags to acquire and transmit locations 24 times a day to a distance of 114Km (line of sight), but provides a surplus of at least 35.40J, which can be used to increase the frequency of position updates or to support alternative communication options such as GPRS. Therefore, this shows the viability of long-term tracking of elephants.
Embargo
Design of a Low-Frequency Linear Motion Testbed for Electromagnetic Kinetic Energy Harvesters in JumboNet
(IEEE, 2017-09-14) Wijesundara, M; Gamage, A; Gokulan, Y; Tapparello, C
Kinetic energy harvesting on animals is an emerging technology that could facilitate real-time monitoring of wild elephants. Real-time monitoring is a proven solution to the Human-Elephant Conflict, a problem that has spread across Asia and Africa. However, when designing electromagnetic harvesters, it is essential to accurately model the voltage produced due to electromagnetic effects. In this paper, we present the design, development and the complete simulation of a simple and low-cost linear motion testbed that estimates the generation of an electromagnetic harvester. We integrated the dynamic non-linear flux linkage across the coil with an analytical model that accurately estimated the motion of the moving magnet. The experimental measurements from the testbed were better than 80% in agreement with the simulation results within the frequency range of 1Hz to 2Hz.
Embargo
Emission Activity Parts Extraction using custom Named Entity Recognition
(IEEE, 2022-12-09) Mannavarasan, M; Gamage, A; Sivarajah, V; Chandrasiri, S
Dental health-related disorders have proliferated worldwide due to the excessive intake of fast food and sugary foods, which was followed by bad oral hygiene practices. The cost of dental examinations may change based on how critical the condition is, regardless of whether they are not regular. For a person, diagnosing an oral health problem, particularly locating the disease’s underlying cause, can be challenging. To properly diagnose and treat such conditions, advanced dental diagnostic techniques may be necessary. By offering convenience and enhancing their oral health knowledge, the system seeks to serve as a prediction tool that regular people can utilize to detect potential tooth illnesses at an early stage. It is encompassed as a mobile application where a Mask R-CNN model is used in the core that accepts a dental radiograph as the input. The trained model will be able to identify diseases related to the bone and teeth. Based on the performance evaluations, the accuracy of the results that are obtained in tooth type, restoration quality, dental caries, and periodontal disease identification falls in the range of 75%-80%.
Embargo
Innovative use of Collaborative Teaching in Conducting a Large Scale Online Synchronous Fresher’s Programming Course
(IEEE, 2021-04-21) Kodagoda, N; Gamage, A; Suriyawansa, K; Jayasinghe, B; Rupasinghe, S; Ganegoda, D; Jayalath, T; Kurrupu, A
The COVID-19 pandemic has forced educationist to come up with innovative solutions in delivering, engaging synchronous online academic modules. An innovative collaborative teaching approach was utilized in delivering programming concepts for freshers. A team of six academics functioned as a resource panel in delivering synchronous online lecture content. These interactive sessions were led by a moderator inquiring the resource panel on topics related to the content of the lecture. This was done in the same spirit on how a panel discussion would be conducted led by a moderator in a conference. Microsoft Teams Live was used in the delivery of the content to an audience of up to 800 students. Delivering a freshers programming course is known to be challenging in face- to-face delivery. A collaborative programming environment was used to engage students in live coding activities during the lectures. Students had opportunities to interact with the resource panel through quizzes, QA and through coding related activities. These lectures also introduced the innovative use of QR codes to get students engagement through a mobile device for the interactive sessions. Results based on a survey shared among the participated students, confirmed the collaborative teaching approach in conducting webinar was more effective over a traditional webinar that is conducted by one person. Interactive programming environment (Repl.it) allowed the resource personal to give feedback on the programs submitted by the students during synchronous sessions conducted. The best practices used in delivering this course can be easily adopted in delivering highly engaging online lectures for other courses.
Embargo
Integrating industrial technologies, tools and practices to the IT curriculum: an innovative course with .NET and java platforms
(acm.org, 2005-10-20) Athauda, R; Kodagoda, N; Wickramaratne, J; Sumathipala, P; Rupasinghe, L; Edirisighe, A; Gamage, A; De Silva, D
Exposure to state-of-art industry technologies, tools and practices by students provide CS/IT graduates highly desirable skills and marketability. A key expectation of the industry from their new cadre is a speedy integration into the business environment resulting in productive work. This usually requires having a sound technological background, a maturity to assess the environment and adapt quickly, and highly-developed soft skills to be productive in a team environment. Incorporating such experience and skills into a CS/IT curriculum is challenging and is still in its infancy stages. We undertook such as an endeavor in integrating .NET into the IT curriculum. Microsoft's .NET platform is becoming increasingly popular in the industry. Incorporating .NET into the undergraduate IT curriculum provides a plethora of skills and increases the employability of our graduates. We integrated .NET without a major revision to the existing curriculum by introducing an optional course in the final year (senior-level) of the IT undergraduate program. In addition to the .NET platform, the course covered the Java platform, which is similar in architecture to .NET. The course emulated an industry-based environment with real-world based assignments, focused on deliverables, used state-of-art IDEs and documentation, and pair programming to create a highly productive environment. The “soft skills” were integrated into the course with a project that implemented a virtual marketplace. Students in groups played different entities in the virtual marketplace and communicated with each other via Web Services. The project provided a virtual business environment and exposure to teamwork, collaboration, competition, negotiating, and creativity skills. Our first offering of the course in semester 1, 2005, attracted 128 students. The course created a highly productive environment throughout the semester. Students completed 7 assignments and the project within the 14-week semester. The initial results are encouraging and provide many insights to CS/IT departments planning to incorporate such courses.
Embargo
Intelligent Timetable Scheduler: A Comparison of Genetic, Graph Coloring, Heuristic and Iterated Local Search Algorithms
(IEEE, 2019-12-05) Ekanayake, T. W; Subasinghe, P; Ragel, S; Gamage, A; Attanayaka, S
A Timetable scheduling is a monotonous task and a problem in an educational institute. This is because many rules and constraints are involved, which can be categorized as hard and soft constraints. Mainly, a university must produce two types of timetables, which are examination, and semester timetables. This paper has reviewed the Exam Timetabling problem with Genetic and Graph Coloring algorithms and the Semester Timetabling problem with Heuristic and Iterated Local Search algorithms. Our aim here is to develop a possible and correct solution for each timetabling problem using the above-mentioned four different approaches.
Embargo
Machine Learning to Aid in the Process of Disease Detection and Management in Soilless Farming
(IEEE, 2022-07-18) Fernando, S. D; Gamage, A; De Silva, D. H
This research aims at enhancing the methods and techniques that are being used in disease detection when it comes to soilless farming. Soilless farming is quite famous among the Sri Lankan farmers farming in urban areas. A mobile application is launched by us and this application is capable of identifying diseases in plants, therefore, farmers do not have to rely on their years of experience to identify the diseases. A novice farmer may struggle to say what is wrong with their plants, while another farmer with many years of experience may say what the disease is with no hesitation. Both those types of farmers benefit from our mobile application equally. The said mobile application consists of four components and each of them focuses on a different service. One of those components is to detect and manage diseases in plant leaves and that component is what this research paper showcases. This particular component allows the user to capture live-images of plant leaves. Then the application processes the captured image to identify if the plant is suffering from a disease. After that, it generates a report with a set of treatments. It further analyses and alerts the user if this disease detected is going to affect the harvest.
Embargo
Mobile Base Solution for Individuals with Limited Knowledge About Cars
(IEEE, 2022-12-09) Nammunige, H; Chamuditha, T; Udara, S; Athapaththu, D; Gamage, A; Gamage, N
Different modes of transportation were discovered by our ancestors from ancient times. Currently, the majority of people choose to purchase a personal automobile for transport needs. However, the vast majority of people are not automobile industry experts. As a result, the majority of people have trouble when recognizing cars. Due to numerous variations of a single vehicle model, even an expert has trouble correctly identifying a certain car model. People must take into account a number of factors before purchasing a specific automobile. Some of crucial factors are service costs and future market prices. Ordinary people require the assistance of a professional when estimating the market price of a car and calculating the cost of servicing a car. Accidents can also occur at any time when driving a car often. In similar circumstances, consumers require the assistance of an insurance agent or a technician to estimate the cost of damage repair. In this study, we provide a way for non-automotive experts to use their smartphones to identify car models, forecast future market prices, determine and forecast servicing costs, and estimate minor damage repair costs. This paper demonstrates how we accomplished aforementioned tasks using YOLO V4, Multiple Linear Regression, Random Forest Classifier and Faster RCNN.
Embargo
A Novel Ranked Emission-Factor Retrieval for Emission Calculation
(IEEE, 2022-11-22) Paskaran, S; Gamage, A; Chandrasiri, S
Emission Factors (EF) selection is a vital task during Carbon Management Systems (CMS) emission calculation. Due to Carbon footprint reduction regulations, there is a demand increase for CMS with better usability and scalability. However, most CMS assumes users know emission technologies well. To circumvent these problems, authors have proposed an approach to building an EF ranking system with a combined scoring approach. It has considered each EF as a document unit and emission activity information provided by the user as the search query. This system uses a linear combination of the Vector Space Model (VSM) and Natural Language Processing (NLP) Word Embedding techniques to rank EF documents for exact and non-exact search queries. This approach's user satisfaction measured with Mean Average Precision (MAP) for “glove-wiki-gigaword-300” at 0.41 linear combination parameter was nearly 30% better than the VSM model and 127% more than the word embedding. In addition, the paper discusses performance metrics such as speed, future EFs scalability, and system resource utilization concerning the solution's overall scalability. This approach can provide better usability and scalable for EF selection tasks compared to single-ranking approaches (VSM or Word Embedding).
Embargo
Planning Marketing Strategies in Small-Scale Business Using Data Analysis
(IEEE, 2022-12-29) Fernando, A.M.P.; Adhikari, A.M.T.T.; Wijesekara, W.H.A.T.K.; Vithanage, T.V.T.I.; Gamage, A; Jayalath, T
The proposed research work develops a system focused on business opportunities to enhance market returns and improve marketing strategies and new strategies by identifying how customers interact with products and their behavior. Existing research efforts attempt to identify and market consumer attraction to products and marketplace in the marketplace. Current research focuses on the challenges of identifying consumer buying patterns and how consumers interact with products, Existing research has not, however, integrated the essential elements into a single system. Consequently, the recommended study has been conducted on a number of significant issues, such as determining the high-value client base and the number of sectors, understanding the purchasing pattern of products that comprise the customers' basket, identifying customer lifetime value, and Customer Trajectory Determination for identifying customer attractive shelf. This system focuses on various machine learning algorithms. Customer segmentation and value analysis using K Mean, Agglomerative, Clustering algorithm, and Arima model. Association rules are generated using the Apriori algorithm for market basket analysis, which is built on the idea that a set of frequently purchased items is a subset of a set of frequently purchased items. Also using RFM analysis to create and prepare our data frame by using BG/NBD model and the Gamma-Gamma model to calculate the customer lifetime value standardization. Using image processing algorithms and retail video analysis algorithms, background reduction technology clearly identifies moving objects/ tracks customer routes using different colors. Based on results from implementation and testing, it was determined that the suggested technique outperformed the use of CCTV to identify consumer behavior and satisfaction with the product in recognizing customer purchasing patterns. The proposed system can identify customers' buying patterns, how customers interact with product...