Other Conference and Symposium Proceedings

Permanent URI for this communityhttps://rda.sliit.lk/handle/123456789/4774

Browse

Now showing 1 - 20 of 115

Embargo
Intelligent Water Quality Monitoring and Prediction System
(Institute of Electrical and Electronics Engineers Inc., 2025) Shiraz, S; Karunasena, K; Mudelige, H; Kumarasinghe, O; Nawinna, D; Perera, J
The paper aims to develop an integrated approach to improve water treatment processes using predictive modeling and SCADA integration in order to meet the specific needs of water purification systems in Sri Lanka. The current systems utilized for this need are outdated since these systems are based on traditional technologies and do not have the means for predictions or real-time data accessibility outside the system. The proposed solution will focus on raw water quality prediction, optimization of chemical usage to bring in efficiency, sustainability, and resource management, ensuring seamless access to all the relevant data required to manage and monitor. In order to achieve this, past data provided by the Meewatura water plant in Sri Lanka, sourced from the Mahawali river, is utilized for the relevant predictions alongside of the data gathered through the SCADA system. The data is not directly accessible since the SCADA system is mainly built for monitoring, and in order to get the data, a MODBUS connection through the PLC is utilized alongside of an IOT device. In addition to the extracted data, past data that was provided by the water plant is also incorporated. The combined data set is utilized for the predictions while continuously improving itself with new data. The present study contributes to the establishment of sustainable and adaptable water treatment frameworks for a wide range of operational needs within the water plants by addressing the gaps in the existing water quality management systems and improving upon them.
Embargo
Blockchain-Based Custody Evidence Management System for Healthcare Forensics
(Institute of Electrical and Electronics Engineers Inc., 2025) Jayasinghe R.D.D.L.K; Sasanka M.W.K.L; Athukorala D.A.S.M; Sandeepani M.A.D; Jayakody, A; Senarathna, A
As digital evidence increasingly growing in significance in healthcare forensics, safeguarding sensitive medical data's confidentiality, integrity, and limited access remains to be an important issue. Existing forensic evidence management systems are subject to data breaches and illegal access since they frequently lack significant privacy-preserving measures. In order to overcome such challenges, this research suggests a Blockchain-Based Custody Evidence Management System for Healthcare Forensics, which combines blockchain technology, machine learning, and encryption methods to improve security, privacy, and accessibility. To ensure accurate and efficient gathering of information, machine learning algorithms are used to extract handwritten and printed text from medical photographs. AES encryption ensures safe storage, while Fully Homomorphic Encryption (FHE) is used for dynamic access level control to protect gathered evidence. Identity verification is made possible via a web-based authentication system that uses Zero-Knowledge Proofs (ZKP) to protect privacy by preventing the disclosure of personal data. By preventing unintended modifications, blockchain technology is used to preserve the custody chain's integrity. Furthermore, machine learning-driven PII detection and masking methods balance the requirement for forensic investigation with privacy compliance by controlling data accessibility according to access entitlements. Based on permitted access levels, the system makes it possible to share safe evidence with law enforcement agencies, such as courts, the police, and other forensic groups. Using blockchain to guarantee data immutability, cryptographic security to restrict access, and artificial intelligence (AI) to safeguard data, this approach enhances the privacy, security, and dependability of handling forensic evidence in medical investigations
Open Access
How E-commerce Succeeds: The Role of Information Systems in Boosting Customer Satisfaction
(Institute of Electrical and Electronics Engineers Inc., 2025) Pathirana, S.L; Pathirana, S.J; Boyagoda, G.S.B; Thalagala, S.M.K; Wisenthige, K; Aluthwala, C
The rapid growth of e-commerce has revolutionized consumer behavior, especially among Millennials and Generation Z, who increasingly rely on online platforms for their purchases. This paper shows the impact of information systems success on individual performance outcomes in e-commerce, focusing on the DeLone and McLean information system success model and its three key dimensions: system quality, information quality, and service quality. A quantitative survey method was employed to gather data from e-commerce users in the western province of Sri Lanka, which is an economically developed region where Millennials and Generation Z are highly engaged with international and local e-commerce platforms. The study uses PLS-SEM to identify that system, information, and service quality significantly increase customer satisfaction and thereby improve individual performance in e-commerce. The study will address the significant research gap in Sri Lanka, where the rapid growth of e-commerce has not been adequately studied in terms of the impact of information and individual outcomes
Embargo
Digital Pathways in Smart Tourism: A Systematic Literature Review on Key Drivers of Travelers' Decision-Making
(Institute of Electrical and Electronics Engineers Inc., 2025) Karunaratna, D; Senevirathne, C; Fernando, K; Peksha, P; Jayasinghe, P; Samarakkody, T
This study systematically investigates how digital components including information reliability, accessibility, cost efficiency, virtual reality and mobile adoption influence travelers' preferences in smart tourism. To do this, the study utilizes the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) framework to aggregate evidence from across the globe - showing both regional technology disparities, and a variety of different methodological approaches. In developed countries, technology integration is quite strong while developing regions face constraints of finance and infrastructure. These tech-centric factors are important to traveler decision-making, as indicated by the review; however, greater coverage of understudied regions is needed. The resulting insights provide the basis for targeted and holistic solutions to improve tourism services for the benefit of policymakers, industry practitioners and academics.
Open Access
A Spatial Study on the Ecological Signatures of Landscapes in Colombo
(Springer Science and Business Media Deutschland GmbH, 2025) Subasinghe J.C; Madhushani T.M.C.I.; Gomes P.I.A
Urbanization is a governing demographic feature and a significant part of global land transformation. According to the United Nations, more than half of the world’s population lives in urban areas. If not studied and managed properly, urbanization can affect negatively its residents, and in Sri Lanka this is about 20%–30% in commercial areas and residential areas. Yet, studies related to exploring functions and status quo of different landuses are rare and rather unfound in Sri Lanka. This study the variations of temperature, humidity, soil moisture, infiltration rate, shrub cover and tree richness with different landuses namely, cemeteries, parks, residential areas and institutes have been investigated to see whether the landuses actually are the landscapes people perceive. It was found that the humidity of land plots with Institutes is significantly higher than all the other landscape types. Interestingly, it was observed that parks and cemeteries possessed high humidity levels while Institutes and Residential areas possessed a comparatively lower humidity level. The soil moisture content and infiltration rates of institutal landscape significantly differed from those of other landscape types. Shrub cover variation between Residential areas and Institutes was insignificant, while shrub cover of all the other landscape types resulted in substantial differences with a significance level of 0.00. The analysis of variation of multiple ecological factors under landscape types depicted that for all the temperatures, the shrubs cover percentage of cemeteries lies higher than the rest of the landscapes. In cemeteries, initially, the shrub cover increased with the humidity and with increments of humidity level, the shrub cover decreased. Overall sense, the Institutional areas depicted relatively adverse liveable conditions, and Cemeteries depicted most favourable conditions, interestingly it was better than Parks. This study gave insights into how these landscapes be best manged and engineering interventions needed in that regard.
Embargo
A BI Approach for Student Engagement and Retention along with Cognitive Load Analysis for Educator
(979-833153098-3, 2025) Algewatta, M. N; Manathunga, K
This research presents a systematic approach to monitoring student engagement, retention, and cognitive load within higher education by integrating Business Intelligence (BI) tools with cognitive load analysis. The proposed framework utilizes a diverse range of data sources -including attendance, academic performance, mental health indicators, demographic variables, and student feedback to generate real-time insights into student behavior patterns. The BI system identified critical trends, such as irregular attendance, declining academic performance, and the influence of demographic factors, enabling educators to identify at-risk students and intervene proactively. Additionally, cognitive load analysis was employed to evaluate the mental demands of course content, categorizing learning objectives in alignment with Bloom's Taxonomy. This allowed for the identification of content that could potentially overwhelm students, facilitating adjustments in instructional complexity. The integration of BI insights with cognitive load data provided a holistic approach that not only enhanced the monitoring of student engagement but also supported the tailoring of instructional content to optimize learning without inducing cognitive overload. The findings suggest that combining BI tools with cognitive load metrics offers a robust framework for both improving student retention and assisting educators in creating a balanced, engaging, and supportive learning environment. This study contributes a practical model for institutions seeking to leverage data-driven insights to promote student success and address the dynamic challenges of modern higher education.
Embargo
Autonomous Water Quality Monitoring: Integrating UWB Ad-Hoc Networks, Sensor Calibration, and Kubernetes Cloud Architecture
(IEEE Computer Society, 2025) Tharindu, K; Abeysinghe, M; Karunarathne, S; Dilshan, K; Primal, D; Jayakody, A
Water quality monitoring plays a critical role in ensuring environmental sustainability and public health. Traditional methods, while accurate, are time-consuming and lack the ability to provide real-time insights. This study proposes a secure, scalable IoT-based solution utilizing autonomous sensor-equipped boats designed to measure pH, turbidity, and temperature in aquatic environments. The boats navigate predefined grid coordinates generated through a Python-based script and communicate data using UWB in a decentralized ad hoc network operating under the AODV routing protocol. Preprocessed sensor data is transmitted to a base station and securely forwarded to a Kubernetes-based cloud infrastructure for real-time processing and visualization. Communication between the base station and cloud services is secured using HTTPS/TLS encryption. Experimental trials confirm reliable navigation, high sensor accuracy, low latency, and robust security. The system remains cloud-agnostic and is compatible with a range of open-source Kubernetes distributions, enabling deployment flexibility across various environments. This research demonstrates an effective, autonomous approach to real-time water quality monitoring, advancing scalable and sustainable environmental management
Embargo
Determining Factors Related to Artificial Intelligence Adoption Among Sri Lankan ICT Service Providers
(Institute of Electrical and Electronics Engineers Inc., 2025) Nizar, Z; Lakshan, J; Wijayarathne, C; Jayasuriya, N; Oshani, W; Rathnapriya, S
This study investigates the factors influencing Artificial Intelligence (AI) adoption among ICT service providers in Sri Lanka, a developing economy where AI integration remains in its early stages. Using the Technology-Organization-Environment (TOE) framework, six key determinants were examined, relative advantage, data quality, top management commitment, employee adaptability, competitive pressure, and external support. Data was collected via a structured questionnaire distributed to 146 ICT service providers, with the sample size determined using G*Power software for statistical robustness. Ordered probit regression was employed to analyze the data, as it offers precise insights for ordinal variables. The findings identify DQ and RA as significant drivers of AI adoption, underscoring the critical role of high-quality data and the operational benefits of AI technologies. However, TMC exhibited a negative impact, highlighting barriers in leadership awareness and alignment with AI strategies. Although CP, ES, and EA were not statistically significant, they demonstrated potential as mediators or moderators in specific contexts. This study bridges a critical gap by providing localized insights into AI adoption challenges and opportunities in Sri Lanka. It emphasizes the importance of data management, leadership commitment, and strategic alignment, offering actionable recommendations for policymakers and industry leaders to enhance competitiveness and digital transformation in the global economy.
Embargo
Multimodal Knowledge Graph for Domain-Specific Intelligence
(Institute of Electrical and Electronics Engineers Inc., 2025) Mohan, K; Munasinghe, M; Bandara, L; Wijesinghe, H; Rathnayake, S; Abeywardhana, L
In the era of information abundance, transforming vast amounts of data into meaningful knowledge remains a critical challenge, especially in domains like medicine, engineering, and education, where visual and multimodal elements play a vital role. Traditional Knowledge Graphs (KGs) excel in organizing structured and textual data but struggle to incorporate multimodal information and implicit relationships, limiting their effectiveness. This paper explores the potential of Multimodal Knowledge Graphs (MMKGs) to address these limitations by integrating text, images, videos, and audio into a unified framework. We investigate how MMKGs enhance knowledge retrieval, comprehension, and interactive learning through advanced techniques, including Natural Language Processing and deep learning. Our findings demonstrate that MMKGs significantly improve knowledge retention and application in specialized fields, offering a foundation for more intuitive and effective domain-specific knowledge ecosystems.
Embargo
Estimation of Switching Overvoltages during Energization of Transmission Lines using Recurrent Neural Networks
(979-833153728-9, 2025) Pelendagamage, S; De Silva, H
Overvoltages frequently pose significant challenges during the energization of transmission lines. During restoration, transmission lines have to be energized from zero voltage to the nominal voltage and the switching of transmission lines is a primary source of these overvoltages. The magnitude and waveform of switching overvoltages are influenced by system parameters, network configuration, and the specific point in the wave cycle at which switching occurs. The ability to estimate peak overvoltages in real-time is crucial for operators, during power system restoration. Traditional methods rely on extensive simulations or empirical formulas, which may not provide the necessary speed or accuracy for operational decisions. It is crucial for operators to ensure that peak overvoltages from switching actions remain within safe limits. This paper introduces a compact long short term memory recurrent neural network (LSTM RNN) based methodology to estimate the peak overvoltages induced during line energization. The developed RNN is trained and tested using extensive simulated data in PSCAD. The results demonstrate that the proposed RNN technique accurately estimates the peak values of switching overvoltages, offering a reliable tool for operators during power system restoration.
Embargo
Enhancing Environmental Awareness for Hard of Hearing Individuals: A Mobile Application Approach
(Springer Science and Business Media Deutschland GmbH, 2025) Dharmasiri, K.G; Rathnasooriya, C.V; Balasuriya, M.K; Yapa, L.N; De Silva, D.I; Thilakarathne, T
This research focuses on developing a mobile application to enhance environmental awareness for deaf and hard of hearing individuals. At its core is an advanced audio classification system using a convolutional neural network model optimized for recognizing environmental sounds. Extensive experimentation identified the best performing convolutional neural network architecture, trained on spectrograms to classify diverse environmental sounds accurately. The model balances accuracy and computational efficiency, making it ideal for real-time mobile deployment. The application includes a user-friendly admin interface, enabling individuals without machine learning expertise to manage and train models, ensuring adaptability to various auditory environments. Leveraging cloud technologies like Amazon Web Services for data storage, processing, and model deployment, the platform provides a scalable solution for safe interaction with surroundings. This empowers users to navigate their environments confidently, enhancing awareness of crucial auditory cues. The study demonstrates the potential of mobile technology to improve inclusivity and environmental consciousness for underserved populations through real-time, tailored sound recognition.
Embargo
"articulearn": An Integrative, AI-Driven Speech Therapy System for Children With Speech Disorders
(Institute of Electrical and Electronics Engineers Inc., 2025) Ranasinghe, K; Zoysa, S.P.D; Annasiwatta, S; Fernando, P; Thelijjagoda, S; Weerathunga, I
"ArticuLearn", a personalized speech therapy system for children with speech sound disorders that integrates advanced machine learning techniques and interactive digital tools to provide targeted intervention across four key domains: phonological disorder detection, fluency disorder identification and intervention, therapy for childhood apraxia of speech, and personalized speech activity filtering for articulation disorders. By leveraging dedicated LSTM-based classifiers and feature extraction techniques such as Mel-frequency cepstral coefficients (MFCCs), this approach automatically identifies specific error types, including phoneme substitutions, omissions, and vowel mispronunciations. In addition, a hierarchical deep learning framework employing attention mechanisms and dynamic time warping is applied to quantify motor planning deficits associated with childhood apraxia of speech, while a reinforcement learning agent adapts therapy prompts based on individual performance. Data were collected from eight children per disorder category along with a normative sample of twenty typically developing children, providing a basis for personalized intervention and progress monitoring. ArticuLearn is designed to complement traditional therapy methods by offering an accessible, scalable solution that supports remote intervention and enhances clinical decision-making. Pilot evaluations suggest that the system can facilitate targeted speech exercises, improve self-monitoring, and foster adaptive learning in young users. This research underscores the potential of combining AI-driven analysis with interactive therapy to transform speech rehabilitation, particularly in resource-limited settings where access to specialized care is challenging.
Embargo
A Game Centric E-Learning Application For Preschoolers
(Institute of Electrical and Electronics Engineers Inc., 2025) Kulasekara D.A.M.N.; Nipun P.G.I.; Dombawela H.M.D.L.B.A; Manilka G.S; Manilka G.S; De Silva D.I.
This research explores the potential of advanced technologies such as pose detection (PD), augmented reality (AR), object detection (OD), and voice recognition (VR) in creating a game-centric e-learning application for preschoolers. The proposed application, Kidstac, integrates cognitive and physical development through interactive activities with real world interaction, addressing gaps in traditional e-learning methods that often neglect physical engagement. The app features real-time feedback mechanisms and structured modules like virtual zoo explorations, exercise games, treasure hunts, and pronunciation activities. Testing results indicate significant improvements in motor skills, knowledge retention, problem-solving abilities, and language proficiency. These findings demonstrate the effectiveness of blending physical and digital learning experiences to enhance early childhood education. The study establishes a foundation for scalable, activity-based learning tools, emphasizing the holistic development of young learners.
Embargo
Adaptive Voice Communication in Emotion-Aware Digital Companions
(Institute of Electrical and Electronics Engineers Inc., 2025) Rathnayake, P; Rathnaweera, C; Jithma, U; Aththanayake, I; Rathnayake, S; Gunaratne, M
This paper presents an adaptive voice communication system for emotion-aware digital companions that dynamically responds to users' affective states through expressive speech and synchronized 3D avatar animation. The system integrates real-time voice input, emotion recognition, and context-aware dialogue generation using GPT-3.5, followed by emotional text-to-speech synthesis via neural TTS. Lip-sync data is generated using phoneme alignment and rendered in sync with the avatar's facial expressions and gestures. To enhance user trust and engagement, the avatar visually mirrors the emotional tone of the speech. A cultural adaptation layer is introduced to align voice output and speech style with Sri Lankan communication norms, including tone, pacing, and formality. Implemented using a Node.js backend and React + Three.js frontend, the system demonstrates strong potential for emotionally intelligent, culturally adaptive AI interactions. This work contributes a modular pipeline for building empathetic voice agents capable of enhancing realism and trust in human-AI communication.
Embargo
Personalized Adaptive System for Enhancing University Student Performance in Sri Lanka
(Institute of Electrical and Electronics Engineers Inc., 2025) Dissanayake, N; Samarakoon, C; Wickramasinghe, D; Pathirana, M; Gamage N.D.U; Attanayaka, B
The growing need for personalized learning strategies has driven the development of data-driven solutions to meet the diverse needs of Sri Lankan university students. A key challenge lies in identifying optimal learning paths that align with individual capabilities, learning styles, and engagement behaviors to improve academic performance. While previous research has explored generalized learning models, these often fail to adapt to the specific demands of individual learners. Traditional strategies lack personalization, resulting in inconsistent learning progress. To address this gap, the research introduces an assistive, data-driven approach that leverages Self-Organizing Maps (SOMs), Adaptive Learning (AL), Content-Based Filtering, Graph Neural Networks (GNNs), and Social Network Analysis (SNA) to create optimized, personalized learning strategies. Clustering algorithms and predictive analysis were used to segment learners and deliver tailored interventions based on their behavior. The proposed system integrates advanced machine learning techniques to enhance student engagement and improve overall academic outcomes through personalized pathways.
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
IoT-Based Smart Hydroponics for Automated Nutrient, Climate, Irrigation, and Health Monitoring
(Institute of Electrical and Electronics Engineers Inc., 2025) Ashik M.A.M; Bogahawatta C.A; Perera M.R.D; Dassanayake D.R.I.P; Jayakody, A; Gamage, N
This study presents HydroNutraLeaf as a selfgoverning hydroponic tower system built with Internet of Things technology to automate the critical aspects of hydroculture farming by uniting water supply management with environmental control and watering systems and plant health monitoring capabilities. The system unites multiple essential components to operate as one unit. The system incorporates an automatic plant disease detection system through real-time image acquisition which uses Convolutional Neural Network (CNN) algorithms and cloud-based warning protocols for classification purposes. An automated system comprising Raspberry Pi actuators, NPK sensors, and machine learning functions delivers nutrients at proper stages during plant growth. A reinforcement learning system directs the management of climate factors including temperature and humidity together with Light Emitting Diode (LED) spectrums to achieve superior yield production and product quality. The system includes a self-operated irrigation system with Electrical Conductivity (EC), potential of hydrogen (pH) regulation features which utilizes SVM-based prediction methods in combination with real-time monitoring to achieve optimum root environment conditions. Users can access a dashboard in Grafana to monitor and control the system by using cloud platforms which include Firebase and AWS. The experimental findings reveal water consumption decreased by 30% along with improved nutritional efficiency reaching 25% and enhanced crop yield reaching 15% with better health performance. The sustainable farming operations and commercial greenhouse implementation benefit from HydroNutraLeaf solution which operates through a scalable model based on data analysis and requires minimal human intervention.
Embargo
Deep Learning Based Sinhala Sign Language Recognition
(Institute of Electrical and Electronics Engineers Inc., 2025) Samarakoon, S.C; Weerasinghe, M
Deaf individuals in Sri Lanka rely primarily on Sinhala Sign Language (SSL) for communication due to hearing impairments. However, effective communication between the Deaf and hearing populations remains challenging due to the limited knowledge of SSL among hearing individuals. This research aims to address this gap by developing an SSL gesture recognition system using computer vision and deep learning techniques. Specifically, the study compares the performance of 3D Convolutional Neural Networks (3D-CNNs) and a hybrid 2D Convolutional Neural Network with Long Short-Term Memory (2D-CNN+LSTM) for classifying short-duration spatiotemporal SSL gestures. Additionally, the research emphasizes reducing computational complexity to ensure efficient operation of the system on low-end devices. These contributions advance the accessibility and practical usability of gesture recognition systems for the Sinhala Sign Language.
Embargo
Project HyperAdapt: An Agent-Based Intelligent Sandbox Design to Deceive and Analyze Sophisticated Malware
(Institute of Electrical and Electronics Engineers Inc., 2025) Perera, S; Dias, S; Vithanage, V; Dilhara, A; Senarathne, A; Siriwardana, D; Liyanapathirana, C
Malware increasingly employs sophisticated evasion techniques to bypass sandbox-based analysis, rendering traditional detection methods ineffective. This research presents Project HyperAdapt: Agent-Based Intelligent Sandbox, a framework that integrates both offensive and defensive machine learning models to enhance malware detection, deception, and behavioral analysis. The offensive RL model generates evasive malware samples, challenging the sandbox, while the defensive models including hybrid evasion detection, GAN-based behavior simulation, and a dynamically adapting RL agent work collectively to improve sandbox resilience. By continuously learning from evasive malware behavior, the defensive RL agent adapts in real-time, strengthening detection capabilities. Experimental results demonstrate that this approach enhances sandbox effectiveness, ensuring long-term adaptability against evolving malware threats.
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.