https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4947 2026-03-29T02:31:15Z https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4957 Title: Enhancing the Communication Experience for the Deaf and Hard-of-Hearing using AI Language Models Authors: De Silva, G N; Bandara, N.G.A.N; Shimra, M.S.F. Abstract: Abstract Effective communication in healthcare is critical for accurate diagnosis and treatment, but deaf and hard-of-hearing (DHH) people experience substantial challenges in medical consultations. This problem grows more significantly in Sri Lanka, since there are few assistive tools accessible for Sinhala-speaking DHH patients. The identified research gap is the absence of accessible, real-time communication tools that enable seamless interaction between healthcare professionals and DHH patients in Sinhala. This research bridges the gap by developing a mobile application that enhances bidirectional communication in medical settings, addressing the user group’s preference for mobile solutions and their need for support in healthcare contexts. The study focuses on creating and evaluating a smartphone application that allows healthcare professionals to speak in Sinhala, which is then transcribed into text. The app provides three contextually relevant responses for the DHH patient, who can choose, modify, and confirm one, which is then turned back into voice for the doctor. Furthermore, the system accepts text-based input from the patient, allowing doctors to answer verbally. The evaluation uses a mixed-method approach, integrating quantitative indicators like mobile application transcription accuracy and answer generating speed with qualitative feedback from interviews with DHH individuals. Usability testing was conducted with Deaf/Hard-of-Hearing (DHH) people and communication facilitators to evaluate accessibility, efficiency, and user satisfaction. The findings show that the application enhances communication clarity and lowers misconceptions during medical consultations. DHH users provided positive feedback, emphasizing the application’s real-time response creation and ease of use. This research contributes to assistive technology for the DHH community in Sri Lanka by providing a realistic answer to a crucial healthcare accessibility issue. 2025-06-30T00:00:00Z https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4956 Title: Enhancing Accessibility for Visually Impaired Individuals Authors: Zahra, M.M.F. Abstract: Abstract This research addresses the challenge of enhancing dining autonomy for visually impaired individuals by designing, developing, and evaluating a mobile application for real-time food recognition. The primary research problem is the lack of accessible and reliable food identification solutions tailored to visually impaired users. To address this, the study explores the integration of advanced object detection and intuitive interaction technologies in a mobile context. The proposed solution employs YOLOv8, a lightweight deep learning model optimized for mobile deployment, paired with a speech-based interface for voice commands and touch gestures. The application enables users to capture a video of their meal and provides spoken descriptions of detected food items through Text-to-Speech (TTS) technology. The backend system leverages a TensorFlow Lite implementation, ensuring low-latency performance on mid-range Android devices. Key system specifications include a Qualcomm Snapdragon 720G processor, 4GB RAM, and Android 11, with backend services running on Firebase for data storage and model updates. Evaluation methods included performance testing and user accessibility studies with visually impaired participants. Results demonstrate high detection accuracy, and positive user feedback. These findings validate the technical reliability and user satisfaction of the solution. This study contributes to the field by demonstrating the feasibility of accessible food recognition technologies through innovative system design and user-centered evaluation. The proposed framework can serve as a foundation for broader applications in accessible technology, improving the quality of life for visually impaired individuals. 2025-04-27T00:00:00Z https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4955 Title: VirExp: Automated Expression and Gestures for Virtual Collaborative Environment Authors: Pupulewatte, P. G. M. N.; Perera, W. U. C. M.; Senadheera, S. M. R. L. A. Abstract: Abstract In an era where virtual collaboration has become integral to professional, educational, and social interactions, the absence of physical expressiveness in digital communication presents a critical challenge. The ”VirExp” study addresses this by developing an innovative real-time pipeline capable of detecting facial expressions and upper body gestures through a standard webcam/inbuilt cameras and representing these as expressive animations via a virtual avatar. This research aims to bridge the expressive gap in virtual environments by translating natural human expressions into dynamic avatar movements, thus enhancing authenticity, expression fidelity, and user engagement in virtual collaboration. The system is designed to answer three core research questions: (1) What specific facial expressions and body gestures convey distinct expressions, and what are the corresponding sequences of skeletal points associated with these gestures? (2) In a real-time skeletal point sequence, how can we identify predefined facial expressions and body gesture patterns within near real-time and express them? (3) To what extent will the suggested solution perform and help users facilitate collaborative interactions in the virtual space? To achieve this, the research employs the Design Science Research Methodology, involving iterative development, rigorous evaluation, and empirical validation. As discussed in Chapter 4, for the technical implementation, skeletal data is captured from 25 participants using the MediaPipe Holistic library, which provides 543 facial and body landmarks. A total of 15,000 frames are collected for each expression. Machine learning models, LSTM, DTW, and Transformers, are trained to recognize expression-linked gesture patterns. As mentioned under Chapter 5, the best-performing model, LSTM-based architecture, achieved 91.67% accuracy. Real-time expression detection is integrated with avatar animation in Unity using Vroid and Mixamo avatars, facilitated by FastAPI for synchronization. User studies involving 30 participants evaluated the system’s performance on both technical metrics and experiential feedback. Expressions, including “High Laugh,” “Subtle Laugh,” “Surprise,” and “Neutral”, were captured and translated into avatar expressions. As discussed in Chapter 5, surveys demonstrated 93.33% user agreement with expression representation accuracy. This research contributes novel skeletal point patterns-based expression representation, introduces a low-cost, accessible framework for real-time avatar expressiveness without relying on sophisticated hardware, and provides empirical evidence of improved collaboration and communication in virtual spaces. Limitations include the focus on a defined set of expressions and exclusion of audio cues, while future work is suggested to expand expression classes, integrate cultural adaptability, and explore broader applications in education, therapy, and immersive metaverse contexts. Ultimately, ”VirExp” redefines how expressions are communicated in digital interactions, offering a technically robust and user-centered solution that transforms avatars from static representations into expression-responsive communicators in virtual collaboration. Key Words: Real-time expression recognition, virtual collaboration, facial expression detection, upper-body gestures, avatar animation, skeletal point tracking, virtual reality, avatar expressiveness, gesture-based communication, virtual environment 2025-06-26T00:00:00Z https://dl.ucsc.cmb.ac.lk/jspui/handle/123456789/4954 Title: Enhancing Adaptive Personalized Learning Interfaces with Generative AI for Individuals with ADHD Authors: Lakshika, P.R.; Gunawardana, L.D.R.N.; Perera, M.D.M.N. Abstract: Abstract Students with Attention Deficit Hyperactivity Disorder (ADHD) face unique challenges in educational environments struggling with their impulsivity, hyperactivity, and inattention issues. Traditional learning platforms often fail to accommodate the specific requirements of ADHD learners, highlighting the need of developing personalized ADHD friendly educational platforms that enhance the learning experience of ADHD individuals by creating an equitable opportunity for them to excel in academics. To bridge this research gap we conducted a study exploring the potential of Generative AI to revolutionize personalized learning interfaces for students with ADHD, aged 18–30, who have prior experience with online learning platforms. This research aims to develop an innovative learning platform tailored for individuals with ADHD improving their encouragement and overall educational outcomes. By applying principles of Human-Computer Interaction (HCI), we compare three distinct learning management systems (LMS): a conventional LMS used by the general population, a manually optimized LMS designed specifically for ADHD users, and an AI-generated LMS customized to ADHD needs through enhanced prompting techniques. This study involves 22 ADHD students whose feedback on each LMS informs our analysis, allowing us to assess how AI-driven customization can adapt interfaces to diverse cognitive and learning styles. The evaluation employs multiple approaches to evaluate the usability and e!ectiveness of each experiment, including the System Usability Scale (SUS), User Experience Questionnaire (UEQ), task analysis, and testing conducted with UI/UX professionals and HCI practitioners, while also evaluating user learning outcomes and engagement levels. The findings reveal contrasting user preferences, with the AI-generated Learning Management System preferred for usability, while the manual system is favoured for e!ectiveness. The findings indicate that although AI-generated user interfaces enhances usability, human-designed user interfaces are essential for educational e!ectiveness. The study promotes hybrid methodologies that combine AI e”ciency with educator-led instructional design, providing practical insights for user interface developers and institutions focused on achieving a balanced user experience and learning outcomes. 2025-06-30T00:00:00Z