Institutional-Repository, University of Moratuwa
Welcome to the University of Moratuwa Digital Repository, which houses postgraduate theses and dissertations, research articles presented at conferences by faculties and departments, university-published journal articles and research publications authored by academic staff. This online repository stores, preserves and distributes the University's scholarly work. This service allows University members to share their research with a larger audience.
 Research Publications |  Thesis & Dissertation |  E- Books |
 UoM Journal Publications |  UoM Conference Proceedings |  Articles published in Scimago's Q1 journals |  UoM Research Reports |  Other Articles authored by UoM staff |
Recent Submissions
item: Conference-Full-text
Enhancing video generation based on text-to-image diffusion models using a multimodal approach
(IEEE, 2025) Sanjula Appuhamy, KD; Miriya Thanthrige, USKP
The rapid evolution of Artificial Intelligence (AI) has revolutionized multimedia creation, particularly through Textto-Image (T2I) diffusion models, which synthesize images from textual descriptions with impressive capabilities. Building on this, enhancing video GIF generation has become a promising frontier. However, the field of Text-to-Video (T2V) GIFs remains underexplored, with limited research addressing this specific application. This synthesis method expands creative expression and finds utility in education, marketing, and entertainment. The lack of focused work highlights both novelty and opportunity for impactful contributions. This study focuses on optimizing T2V GIF generation in terms of computational and parameter efficiency. The resulting model maintains a parameter count below 1 billion, enabling faster training, reduced inference time, lower memory usage, and compatibility with low-end hardware.
Inspired by previous work using 2 × 2 grid diffusion and frame interpolation models, this research proposes a simplified approach using a single Stable Diffusion model. It generates all 16 frames of an animated GIF within a 4×4 grid, eliminating prior post processing steps. Given that the GIF format emphasizes animation over fine detail, this parameter efficient method is well suited.
item: Conference-Full-text
An Integrated sound-based traffic control system (green light) for ambulance prioritization at urban intersections
(IEEE, 2025) Wijesiri, P; Jayasena, S
Urban congestion poses a significant barrier to timely emergency response, particularly for ambulances navigating signalized intersections. This research introduces an integrated, sound-based traffic control system designed to autonomously detect ambulance sirens and grant real-time greenlight priority with minimal disruption to general traffic flow. The system employs low-cost IoT devices installed 10 meters upstream of each lane to continuously monitor ambient sounds. Using a machine learning classifier, a Support Vector Machine (SVM) trained on features such as MFCCs, spectral centroids, and energy metrics, the system achieved 98% accuracy in distinguishing ambulance sirens from urban noise. Once detected, a direction-identification algorithm selects the correct lane by comparing amplitude levels across sensors, while a dynamic timer, based on sound amplitude and speed estimation, calculates the optimal green-light duration. Field trials demonstrated subsecond decision latency and reliable lane identification in all 50 test cases. A convoy mode further supports uninterrupted passage for multiple emergency vehicles by dynamically extending the green phase upon detection of successive sirens. Compared to the conservative fixed 5-second buffer approach, the adaptive system using a 1.5-second dynamic margin reduced unnecessary green light hold time by more than half—cutting average cross-traffic disruption from 5% to just 2% per cycle—while maintaining safe and uninterrupted passage for ambulances across all 50 realworld trials These results affirm the feasibility of using audiobased IoT systems for real-time, intelligent traffic management, especially in resource-constrained urban settings.
item: Conference-Full-text
International Symposium on Earth Resources Management and Environment - ISERME 2025 (Pre-Text)
(Department of Earth Resources Engineering, University of Moratuwa, Sri Lanka, 2025)
item: Conference-Full-text
Evaluation of image-based segmentation algorithms for discontinuity detection
(Department of Earth Resources Engineering, University of Moratuwa, Sri Lanka, 2025) Najeeba, MNF; Silva, KDC; De Silva, KPK; Xavier, SA; Dassanayake, ABN; Thiruchittampalam, S
Accurate detection of discontinuities is critical to determine rock mass features such as block geometry, joint orientation, and potential failure surfaces, which govern structural stability in mining and geotechnical applications. Manual methods of detecting discontinuities are often time-consuming, expose personnel to hazardous situations, and are susceptible to biased judgement. In response, image processing techniques like image segmentation have been increasingly adapted to detect discontinuities from rock outcrop images. This study evaluates the performance of traditional and machine learning segmentation methods to identify a higher accuracy workflow for discontinuity detection and the following methodology was employed: RGB rock outcrop images were manually annotated to establish ground truth masks, pre-processed with noise-reduction filters and then processed using traditional Gradient-based operators, Canny edge detection, thresholding and machine learning approaches, U-net, Holistically-Nested Edge Detection (HED), and Segment Anything Model (SAM). The performance of these methods was quantified by evaluating the Boundary F1 score against the ground truth masks. Among discontinuity-based traditional segmentation methods, on a scale of 0 to 1, Canny edge detection with morphological gradient achieved F1 scores of 0.194 and 0.196, while among similarity-based segmentation methods, dilation of eroded threshold images achieved F1 scores of 0.264 and 0.202. For machine learning methods, SAM outperformed the other methods by achieving F1 scores of 0.752 and 0.632, but caused over-segmentation in highly discontinuous regions. The findings highlight the significance of combining computationally efficient traditional methods with targeted preprocessing for low-resource settings and underscore the trade-off between machine learning accuracy and its infrastructural demands.
item: Conference-Full-text
Study of rock anisotropic effects on mode II fracture toughness at various loading rates
(Department of Earth Resources Engineering, University of Moratuwa, Sri Lanka, 2025) Ichikawa, K; Min, G; Fukuda, D; Kawasaki, S; Kodama, J; Cho, S
Fracture toughness is an important parameter for evaluating the resistance of a material to crack initiation and propagation. Although Mode I and Mode II fracture toughness have been studied under quasi-static loading, their behavior under dynamic loading remains insufficiently understood. Previous research shows that compressive and tensile strength, as well as Mode I fracture toughness, are influenced by loading rate and often result in different fracture patterns. Based on this, it is expected that Mode II fracture toughness may also be sensitive to loading rate. Additionally, rock anisotropy, which affects crack propagation, may influence fracture behavior under varying loading conditions. This study first used Finite Element Method simulations with the J integral to evaluate geometry-related factors. Then, Mode II fracture toughness tests were conducted at different loading rates using the Short Core in Compression method. A servo-controlled hydraulic system and a Split Hopkinson Pressure Bar were used to apply quasi-static and dynamic loading, respectively. The effects of loading rate and anisotropy on Mode II fracture toughness and crack propagation were examined.