Browsing by Author "Krishara, J"

Now showing 1 - 15 of 15

Embargo
AI Powered Integrated Code Repository Analyzer for Efficient Developer Workflow
(Institute of Electrical and Electronics Engineers Inc., 2025) Akalanka, I; Silva, S.D; Ganeshalingam, M; Abeykoon, A; Wijendra, D; Krishara, J
Transitioning between new and legacy codebases in diverse project environments poses significant challenges for developers, especially with traditional Knowledge Transfer (KT) methods, which are often resource intensive and prone to obsolescence. These limitations hinder the Software Development Life Cycle (SDLC), particularly in fast-paced industrial settings. This research introduces an AI-driven automation solution that leverages large language models (LLMs) and advanced artificial intelligence technologies to address critical gaps in technical knowledge transfer, with a focus on modern software frameworks. The proposed system reduces development costs, improves team performance, and accelerates adaptation to complex codebases. Key features include a documentation generation tool that cuts manual effort by up to 90%, with an average generation time of 6.8 minutes. Additionally, a virtual knowledge transfer assistant enhances onboarding efficiency, potentially reducing senior developer involvement by 50-60%. The system also includes an automated diagram generator that achieves 97% validation accuracy and a code smell detection tool with 71% accuracy, resulting in better code quality assessments. These findings demonstrate the effectiveness of AI-driven automation in improving developer productivity, streamlining onboarding processes, and optimizing software development workflow
Embargo
AI-Driven Vehicle Valuation and Market Trend Analysis for Sri Lanka's Automotive Sector
(Institute of Electrical and Electronics Engineers Inc., 2025) De Silva K.P.N.T.; Shehan H.A.; Jayawardhane A.S; Premarathne A.P.S.; Krishara, J; Wijendra, D.R
The automotive sector in Sri Lanka faces challenges in vehicle valuation accuracy and market trend analysis due to fluctuating prices, varying vehicle conditions, and environmental concerns. This paper presents an AI-driven vehicle valuation system integrating machine learning models for automated vehicle identification, damage detection, market trend analysis, and environmental sustainability assessments. Using deep learning techniques such as Convolutional Neural Networks (CNNs) and time-series models like Long Short-Term Memory (LSTM), the system delivers accurate valuation and market trend insights. Experimental results demonstrate 9 2% accuracy in damage classification and a mean absolute error (MAE) of 5.3% in repair cost estimation, supporting informed decision-making. This research bridges gaps in valuation transparency and sustainability in emerging automotive markets.
Embargo
An Integrated Deep Learning Framework for Early Detection of Vision Disorders
(Institute of Electrical and Electronics Engineers Inc., 2025) Jayathilaka, S; Balaruban, D; Kumanayake, I; Elladeniya, A; Wijendra, D; Krishara, J; De Silva, M
Vision impairment due to retinal diseases like Diabetic Retinopathy (DR), Age-Related Macular Degeneration (AMD), Glaucoma, and Retinal Vein Occlusion (RVO) poses a significant health challenge in Sri Lanka, where these conditions are leading causes of blindness. This research presents a novel multi-disease prediction system leveraging advanced deep learning techniques for early detection of DR, AMD, Glaucoma, and RVO. The study utilized publicly available datasets, including retinal fundus images from repositories such as RFMiD, IDRiD, APTOS validated by medical professionals to ensure diagnostic reliability. These images were preprocessed and augmented to train robust convolutional neural network (CNN) models tailored to each disease. The predictive models were developed and optimized using hybrid architectures, integrating attention mechanisms and feature fusion for enhanced performance. This approach achieved high accuracies 93% for DR, 92% for AMD, 94% for Glaucoma, and 94% for RVO demonstrating robustness and consistency across diverse retinal conditions. To validate real-world applicability, the models underwent further testing in clinical settings using a Sri Lankan dataset, reflecting local disease prevalence and imaging conditions. By combining validated public data with clinical testing, this scalable system supports ophthalmologists in early diagnosis, reducing diagnostic delays and improving patient outcomes. This work offers a reliable, innovative solution to mitigate the burden of blindness in Sri Lanka and beyond.
Embargo
Automated Sinhala Voice Assistant to Manage Tasks using Natural Language Processing - ශ්‍රී Voice
(IEEE, 2022-12-26) Senarathne, K.H.I.R.; Nirash, J.M.I.; Herath, H.M.C.P.; Bandara, V.D.; Wijendra, D; Krishara, J
Voice assistants are programs on digital devices that listen and respond to verbal commands. In this dynamic world, users can use these voice assistants to manage daily tasks, plan their day, get answers to problems, and for entertainment purpose. Most of existing voice assistant applications functioned using the English language. Since Sinhala is the native language in Sri Lanka, it is not recognized internationally as well as within Sri Lanka for technical applications. As Sri Lankans are more inclined to use the Sinhala language, it is expected to develop this for the benefit of all the Sri Lankans despite their age and to use their native language through a technical application. Furthermore, the lack of English knowledge will lead to the decline of Information technology literacy. This project expects to take the Sinhala language forward to a standard where it is recognized locally and internationally. Thus, building a mobile application that supports Sinhala voice commands will solve the above-mentioned issues. It will facilitate the individuals to do daily activities efficiently and effortlessly within a less time. Machine Learning and Natural Language Processing are the man technologies used in this project. A computer vision-based algorithm from face detection technology is trained. This application is capable of answering questions and following the instructions for daily tasks and also can be used for entertainment. Apart from the mentioned specialty, it also supports and responds to the Sinhala language that will be shown off on a display.
Embargo
A Comprehensive Mobile Platform for Fostering Communication, Literacy, Numeracy, and Emotion Understanding in Children with ASD
(IEEE, 2024-07-25) Bandara, T.W.M.I.P.S; Deshan, M.A.D.; Prasanth, P.; Nadeera, M.S.; Krishara, J
This study presents SIPNENA, a novel mobile application designed to aid the learning and communication development of Sinhala-speaking autistic children aged six, particularly in rural areas of Sri Lanka. It offers a unique approach to teaching challenging subjects like English and Mathematics, tailored to the specific needs of children with Autism Spectrum Disorder (ASD). The application integrates interactive methodologies and gamification elements to facilitate better communication, understanding, and engagement. Additionally, it incorporates real-time emotion recognition features to monitor and respond to children's emotional states during learning activities. This research evaluates SIPNENA's effectiveness in improving communication abilities, academic skills, and emotion understanding among autistic children. The findings indicate promising results in catering to the unique educational needs of this target population, particularly in under-resourced rural regions, where specialized interventions are often scarce.
Embargo
Context-Aware Behavior-Driven Pipeline Generation
(Institute of Electrical and Electronics Engineers Inc., 2025) Gunathilaka, P; Senadheera, D; Perara, S; Gunawardana, C; Thelijjagoda, S; Krishara, J
Enterprise networks increasingly rely on cloud platforms, remote collaboration tools, and real-time communication, placing high demands on bandwidth availability and responsiveness. Static bandwidth allocation approaches often fail to adapt to dynamic traffic conditions, leading to congestion, inefficiency, and degraded Quality of Service (QoS) for critical services such as VoIP and video conferencing. This research introduces a novel real-time bandwidth allocation system that integrates Deep Packet Inspection (DPI), supervised machine learning, and Linux traffic control (tc). Unlike prior solutions that focus only on classification or simulation, our system actively enforces bandwidth policies based on live predictions. Traffic is captured and analyzed in the WAN, while adaptive policies are deployed in the LAN. A web dashboard offers real-time traffic and bandwidth visibility. The proposed system addresses realworld enterprise challenges by enabling intelligent, responsive bandwidth management without requiring costly infrastructure changes, achieving measurable improvements in latency, throughput, and application-level prioritization
Embargo
Enhancing the Performance of Supply Chain using Artificial Intelligence
(Institute of Electrical and Electronics Engineers Inc., 2025) Wijedasa, S; Gnanathilake, K; Alahakoon, T; Warunika, R; Krishara, J; Tissera, W
Optimizing warehouse operations is essential to meet dynamic customer demands while maintaining efficiency in the rapidly changing supply chain landscape. Using four key components, this research presents a comprehensive AI-based approach to improve supply chain management performance. The first component uses Long Short-Term Memory (LSTM) networks to predict demand and returns, allowing for accurate forecasting of product demand and returns based on historical sales data. The second component uses Q-learning, a Reinforcement Learning (RL) technique that optimizes the scheduling of product replenishments by prioritizing critical stock shortages based on inventory levels and predicted demand. The third component analyzes customer purchasing patterns using FP Growth and clustering algorithms to analyze customer buying patterns, strategically placing items in aisles to reduce selection time and improve picking efficiency. The final component involves customer churn prediction using machine learning techniques to identify at-risk customers and facilitate proactive retention strategies. To bridge the gap between complex AI models and practical warehouse operations, a web-based application named 'OptiFlow AI' has been developed. This platform provides warehouse workers with user-friendly interfaces to access demand forecasts, replenishment priorities, optimized product placements, and customer retention insights. The proposed system significantly enhances operational efficiency, reduces time delays, and improves customer satisfaction, contributing to a more resilient and intelligent supply chain ecosystem.
Embargo
Face Skin Disease Detection and Community based Doctor Recommendation System
(IEEE, 2022-12-09) Udara, M.A.A.; Wimalki Dilshani, D.G.; Mahalekam, M.S.W.; Wickramaarachchi, V.Y.; Krishara, J; Wijendra, D
In our country, skin diseases are more common than other diseases because of the climate. Skin diseases are occurring almost on all groups of ages among people. It is one of the most common types of diseases where some can be painful, and some can cause fatal to human life. The delay of the disease detection, difficulties of identify the infected area, Ignorance of the spread of the disease and treatments may threat to the patient’s life. Most of the time this process is performed manually which can lead to human errors and takes days for providing the results. This paper reports a smart solution that assists the patients by detecting the disease, identify the current infected area of the disease, recommend best doctors, provide community-based prevention guidelines, and predict the future risk. Also due to this economic crisis, we suggest that it’s much easier if the patient can do these skin check-ups systematically to continuously monitor and detect skin disease to get proper medical attention. As treatment procedures can be different from each doctor and impact will be different, we are working on community-based platform where we can get patients’ reviews about doctors and preventive guidelines. Depending on the performance evaluations, the results obtained from the proposed method for disease identifications are in the range of 90% - 95% of accuracy.
Open Access
Intelligent Systems for Comprehensive Dog Management
(Association for Computing Machinery, 2025-06-28) Katipearachchi, M.E; Sachethana, O; Gunawardena, G. N.A; Ruwanara, D.C; Krishara, J; Kasthurirathna, D
In recent years, the integration of advanced technologies with canine welfare has gained significant attention, leading to the development of comprehensive platforms for dog management. The "Research Pooch-Paw"initiative addresses the multifaceted needs of dog owners and stray dog populations through an innovative platform that incorporates machine learning, wearable sensors, and real-time data processing. The platform facilitates early disease detection, behaviour analysis, and health monitoring using IoT-enabled devices, and provides personalized care guidance. Additionally, it includes features for stray dog identification and emergency response using deep learning algorithms and image processing techniques. The research underscores the potential of leveraging modern technology to enhance the quality of life for dogs and improve the effectiveness of canine welfare strategies.
Embargo
IoT-Based Solution for Fish Disease Detection and Controlling a Fish Tank Through a Mobile Application
(IEEE, 2024-04-05) Bodaragama, B.D.T; Miyurangana, E.H.A.D.M; Jayakod, Y.T.W.S.L; Vipulasiri, D.M.H.D; Rajapaksha, U. U. S; Krishara, J
This research project seeks to enhance fish tank management and improve the well-being of aquatic life by leveraging modern technological solutions. It focuses on four key areas: monitoring water quality, detecting fish diseases, preventing algae growth, and developing an automatic fish feeder with remote control capabilities. The project’s first goal is to establish a comprehensive water quality monitoring and control system that predicts future water conditions, continuously assesses key parameters, and provides real-time data to users for proactive interventions. Additionally, the research project aims to develop an image-processing-based mobile application for early detection of fish diseases, eliminating the need for manual inspection and improving overall fish health management. The project also involves the creation of a mobile app to predict and prevent algae growth by analyzing factors like lighting, nutrient levels, and water flow, providing personalized recommendations for algae control. Lastly, an automatic fish feeder with remote control capabilities will be designed, allowing fish owners to schedule and adjust feeding times and portion sizes through a mobile app. This innovative approach ensures fish receive consistent and appropriate nutrition even when owners are away from home.
Embargo
A Mobile-Based Screening and Refinement System to Identify the Risk of Dyscalculia and Dysgraphia Learning Disabilities in Primary School Students
(IEEE, 2021-08-11) Hewapathirana, C; Abeysinghe, K; Maheshani, P; Liyanage, P; Krishara, J; Thelijjagoda, S
Learning Disability is a condition that has a direct effect on the brain and there is no cure or any identified medical treatments. Most of these cases remain undiagnosed due to the lack of awareness from their parents and teachers in underdeveloped countries like Sri Lanka. Mobile application-based solution ‘Nana Shilpa’ was developed for the screening and intervention processes for the specific Learning Disabilities which are Verbal and Lexical Dyscalculia, Operational and Practognostic Dyscalculia, Letter Level Dysgraphia and Numeric Dysgraphia. Deep Learning with Machine Learning techniques is used in the screening process to provide a better solution. To detect the written letters/numbers, trained Convolutional Neural Networks (CNN) achieved the accuracy of 92%, 99%, 99% for Verbal and Lexical Dyscalculia, Letter Level Dysgraphia and Number Dysgraphia respectively. The Machine Learning algorithms used for screening processes are Support Vector Machine (SVM) and Random Forest (RF). In the machine learning models, it is achieved the accuracy of 98%, 97% for Operational and Practognostic Dyscalculia and Number Dysgraphia respectively. In Sri Lanka, this has been recognized as an acceptable solution for screening and intervention via a mobile-based application for above mentioned four variants of learning disability conditions which are developed based on the gaming environment.
Embargo
Revolutionizing Tamil Language Analysis: A Natural Language Processing Model Development Approach
(IEEE, 2024-07-25) Ravichandira, G; Sivabaskaran, V; Uthayakumar, T; Vyravanathan, S; Krishara, J; Rajendran, K
This study proposes a web-based platform utilizing Natural Language Processing (NLP) techniques to identify and rectify spelling and grammar errors in Tamil, a language with intricate nuances. Users can input Tamil text, which undergoes automated scrutiny for linguistic inaccuracies. Additionally, the research delves into contextual text summarization and real-time transcription of spoken Tamil. The overarching aim is to devise a holistic solution amalgamating various components to facilitate the detection and rectification of Tamil spelling and grammatical errors. The envisioned subgoals encompass a spell-checking tool capable of identifying misspelled words and suggesting appropriate replacements based on context, a grammar correction feature adept at identifying and rectifying grammatical inaccuracies while accommodating the unique grammatical structures of Tamil, a summarization component adept at condensing paragraphs while retaining core concepts, and a transcription feature enabling the real-time conversion of spoken Tamil into accurate text. By addressing the complexities of the Tamil language, this research endeavor seeks to contribute to the expansion of language processing tools. The ultimate objective is to empower users with the means to detect and rectify errors while enhancing their proficiency in spoken Tamil. This synthesis of components represents a significant stride towards the development of a comprehensive web-based platform for identifying and rectifying Tamil spelling and grammar errors.
Embargo
Sell-Net: Mobile Based Solution to Strengthen the Seller-Buyer Relationship of Small-Scale Businesses in Sri Lanka
(IEEE, 2022-01-21) Wickramarathne, N; Siriwardana, U; Karunarathna, A; Jeremiah, P; Thelijjagoda, S; Krishara, J
As a developing country, Sri Lanka is getting a greater number of people who are using an online application and online purchase for various kinds of items nowadays. The number of users is highly increasing day by day. These online applications and their usages have a major role in the economy of Sri Lanka. The usage of online purchasing is widely observed in Sri Lanka as well it has been observed that most of the delivery persons and small-scale business people are facing many challenges while using these applications in practical ways. This is an application that connects small-scale business sellers directly to the buyer on an online platform. Goods Delivery vehicles are widely spread in Sri Lanka with traveling their paths. Most of the time they traveled with some extra free spaces. Authors are trying to manage these vehicles to give up an extra service and get extra income with proper management. It tries to improve the service to get more people an advantage and build up their businesses. This app aims at three main user areas: seller, delivery persons, and buyer. The buyer can place the order for these goods through the online platform and then the system calculates and find the nearest driver to the seller and choose the driver. From using the app driver sees the best path to the seller. Also, there's an algorithm to calculate ordered goods type, quantity, size and select which vehicle is suited for delivering these goods. The delivery person delivers those goods to the buyer by their path. The System tracks down until the buyer received their goods. and also detecting the fake review also if buyer using this application never get unsatisfied for the practical usage. The outcome of this project will help to develop a reliable seller-buyer relationship for small-scale business parties and improve their business activities.
Embargo
A Smart Waste Disposal System: To Encourage Proper Waste Disposal
(IEEE, 2022-12-09) Alwis, D; Munasinghe, P; Rajapaksha, S; Ranawaka, B; Krishara, J; Tissera, W
Waste disposal is one of the most important industries in the world. If not maintained properly it would lead to the destruction of the environment. Improper waste disposal is becoming a critical issue in Sri Lanka and the lack of waste segregation, inadequate waste collection methods, the lack of support for waste management from the public are among the root causes of the problem. As a solution we propose an IoT-based solid waste management system that allows garbage bin monitoring, routing of garbage collector trucks, a prediction model and a point rewarding system. As the end result of this research the following prototypes was built; a prototype model of a smart bin with the capabilities of opening and closing by itself and detecting the waste level of the bin, a prototype mobile application for garbage collectors which delivers analysed data on truck position and ensures timeliness, a prototype mobile application for the public which receives the weight and type of solid waste discarded as an input and calculate reward points to encourage the public in proper waste disposal, a prototype web application which delivers statistical data for detailed reports and a prediction model which predicts the amount of waste to be collected in the coming month using machine learning. This is a low-cost IoT-based solution that uses existing resources to handle the massive amounts of garbage collected each day.
Embargo
Throat AI - An Intelligent System For Detecting Foreign Objects In Lateral Neck X-Ray Images
(Institute of Electrical and Electronics Engineers Inc., 2025) Baddewithana, P; Krishara, J; Yapa, K
Foreign Object ingestion is a commonly encountered medical condition within the Ear, Nose, and Throat clinical domain. Timely and accurate detection of such objects is vital, as it often guides the need for surgical intervention. Among the available imaging techniques, lateral neck X-rays are the most widely used radiographs to visualize and assess the presence of FOs in the throat. However, manual interpretation of these images can be time-consuming and subject to human error, potentially leading to misdiagnosis or delayed treatment. This research presents a deep learning-based software solution, deployable via web and mobile platforms, aimed at assisting medical professionals with the automated detection of FOs in lateral neck X-rays. The system leverages state-of-the-art YOLO object detection models, specifically evaluating novel versions such as YOLO-NAS-s, YOLOv11s, and YOLOv8s-OBB to ensure high detection accuracy and deployment efficiency. The best-performing model, YOLO-NAS-s, achieved a validation accuracy of 96.3%. For deployment, the model was hosted on the Roboflow platform and accessed via a FastAPI-based middleware server. Performance evaluation showed an average inference time of approximately 2 seconds and a memory footprint of around 100 MB on standard computing hardware, demonstrating its suitability for integration into resource-constrained clinical environments. This setup highlights the system's lightweight design and real-world applicability. Training, evaluation, and testing of the deep learning models were conducted using a dataset curated from public local healthcare institutions and online medical imaging repositories.