DCU-Final-Year-Projects-Booklet-2025

40 24. WasteMaterial Classification Using Computer Vision This project’s purpose is to utilise deep learning and computer vision to to classify images of rubbish in landfill. Deep learning architectures are used in tandemwith various data augmentation methods for image classification. The project also aims to determine the most appropriate deep learning architectures and data augmentation techniques for this application. This project attempts to provide a solution to the ever growingmountains of rubbish in landfills, that have negative impacts on our environment. In keeping with the theme of sustainability, this project is further optimised to run with minimal computational power by explicitly balancing both the accuracy and computational efficacy of the models evaluated. Student Programme Data Science Project Area Artificial Intelligence, Computer Vision Project Technology Python, Machine Learning Student Name(s) Orlaith Quinn  |  Caoimhe Duignan Email orlaith.quinn22@mail.dcu.ie   |  caoimhe.duignan8@mail.dcu.ie Supervisor Dr Ellen Rushe 25. MirAI: Local AI Assistant MirAI is a smart assistant developed to offer an open-source alternative to cloud-based assistants like Google Home and Amazon Alexa. The project targets users with privacy concerns about common cloud-based smart assistants, providing themwith control over their data and a framework that allows for customisation to suit individual preferences. MirAI is designed to be extensible, enabling community- driven customisation and interaction with user-specific services. Users can add customAPI calls with corresponding activation phrases, such as linking a weather API to the phrase, ‘What’s the weather like in Dublin?’ The backend can be run on Linux or Windows hardware that meets or exceeds the minimum hardware requirements Student Programme Computer Applications Project Area Artificial Intelligence, Internet of Things, Natural Language Processing Project Technology CSS, Docker, HTML5, JavaScript, MySQL, Python, React JS, Flask Student Name(s) Georgijs Pitkevics  |  Chee Hin Choa Email georgijs.pitkevics2@mail.dcu.ie   |  chee.choa2@mail.dcu.ie Supervisor Prof Gareth Jones 26. PredictingHerd-Level Bovine Tuberculosis Breakdowns in Ireland UsingMachine Learning The objective of this project is to tackle the challenge of reducing the spread of bovine tuberculosis (bTB) in Ireland by predicting its transmission and identifying at-risk cattle herds based on herd-level characteristics and cattle movement patterns. We have employed graph-based techniques tomodel and extract insights into disease transmission, alongside traditional machine learning algorithms such as Random Forest, XGBoost, Logistic Regression, and Neural Networks. Data for this project has been sourced from the Department of Agriculture, Food and theMarine. By providing actionable insights, this project aims to support decision-making and contribute to controlling the recent spread of bTB. Student Programme Data Science Project Area Artificial Intelligence, DataMining, Databases Project Technology Python, SQL, Machine Learning, Neo4j Student Name(s) Vrinda Kallu  |  BartoszWalkowiak Email vrinda.kallu2@mail.dcu.ie   |  bartosz.walkowiak2@mail.dcu.ie Supervisor Dr AndrewMcCarren