Browsing by Author "Wijesinghe, N"

Filter results by typing the first few letters
Now showing 1 - 4 of 4
  • Thumbnail Image
    item: Article-Full-text
    AI-powered smart recycling: turning plastic trash into treasure
    (University of Moratuwa, 2023) Kristombu, S; Thilakarathne, BS; Perera, S; Mendis, P; Ruwanpathirana, G; Rohanawansha, H; Wijesinghe, N; Mallikarachchi, C; Weerasinghe, P; Herath, S
    In a world grappling with environmental challenges posed by plastic waste, innovative solutions are emerging to address the pressing issue of plastic recycling. Among these solutions, Smart AI-enabled automation and Upcycling stand out as promising technologies that offer the potential to revolutionize the way we handle and repurpose plastics. These technologies harness the power of artificial intelligence (AI) and automation to streamline the recycling process and transform discarded plastic materials into valuable products.
  • Thumbnail Image
    item: Article-Full-text
    Anomaly detection in image streams with explainable AI
    (University of Moratuwa, 2023) Wijesinghe, N; Perera, R; Sellahewa, N; Talagala, P
    We define an anomaly as an unlikely occurrence that deviates from a typical behavior [1]. An anomaly could be a defect in a production line, sudden stock market fluctuations or natural disasters such as deforestation, volcanic eruptions, or floods [2] [3]. The assistance of an intelligent system to identify such disturbances would be very beneficial to initiate methods to prevent such situations in the early stages. This study forwards an AI based anomaly detection system and its testing stages primarily focused on the detection of deforestation, where when deforestation occurs, it shows an anomalous scenario which deviates from the typical sights of lush green forests.
  • Thumbnail Image
    item: Conference-Full-text
    A Bus route-efficiency analysis using stochastic frontier model
    (Sri Lanka Society of Transport and Logistics, 2024) Wijesinghe, N; Mudunna, R.; Herath, D; Wickramasinghe, V.; Gunaruwan, T. L.
    This research analyzes the efficiency of bus routes in Kandy, Sri Lanka, utilizing a Stochastic Frontier Model (SFM). The primary objective is to identify inefficiencies and provide recommendations for optimizing bus operations. The data collection was done through collecting relevant documents maintained at bus depots. These documents included records of revenue, number of buses, distance covered, fare structure, load factors, route length, and bus capacities. The analysis reveals that over 60% of routes cluster around a moderate efficiency range, indicating potential for improvement. Notably, a significant number of routes (potentially 24 or more) achieve efficiency scores exceeding 90%, serving as exemplars of efficiency. Conversely, some routes score as low as 10%, necessitating further investigation to address inefficiencies. The SFM results, with an R-squared value of 0.84, indicate the model explains approximately 84% of the variation in revenue per kilometer. Key recommendations include optimizing bus schedules to increase load factors, adjusting trip lengths to balance operational costs and passenger demand, and re-evaluating bus allocations to ensure resource efficiency. These findings provide actionable insights for enhancing the efficiency of bus routes in Kandy, contributing to improved public transportation services and resource allocation.
  • Thumbnail Image
    item: Conference-Full-text
    Early identification of deforestation using anomaly detection
    (Information Technology Research Unit, Faculty of Information Technology, University of Moratuwa., 2023-12-07) Wijesinghe, N; Perera, R; Sellahewa, N; Talagala, PD; Piyatilake, ITS; Thalagala, PD; Ganegoda, GU; Thanuja, ALARR; Dharmarathna, P
    Research involving anomaly detection in image streams has seen growth through the years, given the proliferation of high-quality image data in various applications. One such application that is in urgent need of attention is deforestation. Detecting anomalies in this context, however, remains challenging due to the irregular and low-probability nature of deforestation events. This study introduces two anomaly detection frameworks utilizing machine learning and deep learning for the early detection of deforestation activities in image streams. Furthermore, Explainable AI was used to explain the black box models of the deep learning-based anomaly detection framework. The class imbalance problem, the inter-dependency between the images with time, the lack of available labelled images, a datadriven anomalous threshold, and the trade-off of accuracy while increasing interpretability in the black box optimization methods are some key aspects considered in the model-building process. Our novel framework for anomaly detection in image streams underwent rigorous evaluation using a range of datasets that included synthetic and real-world data, notably datasets related to Amazon’s forest coverage. The objective of this evaluation was to detect occurrences of deforestation in the Amazon. Several metrics were used to evaluate the performance of the proposed framework.