2012 | 2013 | 2014 | 2015 | 2016 | 2017 | 2018 | 2019 | 2020 | 2021 | 2022 | 2023 | 2024
[PDF/Print]

Projects from 2023

Student	Title	Supervisor	Moderator	Module
Abbie Price	Design and Evaluation of a Framework for Secure and Accessible Authentication for Individuals with Disabilities of Dexterity	Fernando Loizides	Irena Spasic	202223-CM3203-RP
We would like to investigate how people with motor disabilities are able to use different modalities in order to authenticate themselves in a situation such as a pin for a credit card or entering their details at an ATM machine.[more]
Abdullah Alqarni	IoT security (multiple projects)	George Theodorakopoulos	Andrew Hood	2223-CMT400
IoT devices (smart lightbulbs, smart plugs, smart cameras, etc) are entering the market without much regard for security. A project in this area will focus on designing and implementing attacks against IoT devices, or on defences that mitigate existing attacks. This project requires either hands-on experience with or willingness to learn network traffic sniffing and pentesting at the level of CMT121.[more]
Abid Gafoor	Review of current publicly available datasets for training AI	Eirini S Anthi	Yulia Cherdantseva	2022-CMT400
This project will review the current literature within Cybersecurity and identify all the available datasets that can be used for training AIs. The main goal of the project is to categorise these datasets and determine their fitness for usage. To achieve this: i) The project will require the design and creation of a fitness method and categorization of AI data sets. ii) The project will also investigate if these data sets are universally useful for BlueTeam AIs and RedTeam AIs? If not, what w...[more]
Abinash N Ramesh	Risk Quantification and Analysis of Malicious IoT Network Traffic	Amir Javed	Theodoros Spyridopoulos	2223-CM3203
Stage 1 Identify a a publicly available dataset that has a variety of IoT attacks. Stage 2 Preprocess the dataset, then based on the attack you will calculate risk for each attack. Annotate the dataset based on the risk score. Propose a risk quantification formula. Stage 3 Use ML/Statistical model to predict risk. Stage 4 (desireable) create a dashboard to demonstrate the working of your project. Demonstrate continuous risk.[more]
Adam Ghebrenegus	Web testbed for programming projects	Sylwia Polberg	Hiroyuki Kido	2223-CM3203
The aim of this project is to create a web application that would allow students to submit their coded Python and/or Java projects and execute them against the settings defined by the teacher. The student uploads need to be safe and secure, and the testing files provided by the teacher as well as possible settings need to be safely hidden and inaccessible to students.[more]
Alan Dunster	Developing a web application to Visualise IoT Data in the Abacws Building	Charith Perera	Federico Liberatore	2021PT-CMT404
In this project, you will be developing something like this https://3d.usb.urbanobservatory.ac.uk/ You will develop Javascript skills and some IoT skills as well. You are free to use whatever the technology and frameworks in this project. Some useful pointers and recommendations are: https://threejs.org/ See this: https://threejs.org/examples/#webgl_loader_ifc You do not start from scratch. You are welcome to utilise as much open source and already developed code based as possible and extend t...[more]
Alec Cook	Generating Commentary of Football Matches Using Natural Language Processing	Oktay Karakus	Hantao Liu	2223Place-CMT403
Generating Commentary of Sporting Events Using Natural Language Processing I would like to create a rudimentary model that takes in raw sporting data (focusing on football matches) and produces commentary about the events in a match. I’ve been a big fan of football from a young age and have always had a slightly unhealthy obsession with the football manager game series. I was always intrigued by how the game simulates fake football matches and would like to attempt to replicate it in the real...[more]
Alejandro Arroyo	Further Implementation of an Mobile Application to Support Caregiver's Practices to Enhance Health and Wellbeing of Infants in Peru	Nervo Verdezoto Dias	Omer F Rana	2223-CM3203
Prerequisites: -Basic knowledge of Human-Computer Interaction, Design Thinking and User Experience (e.g., user-centered and/or participatory design methods) -Experience developing mobile and/or web applications in several platforms (e.g., Android, HTML5 & Javascript, etc.) -Desirable experience with Spanish (not mandatory) A PhD student together with 2 bachelor students are developing a mobile application to support parents managing the health and wellbeing of their infants as well as the f...[more]
Alex Hayman	Droidloader: Generating New Datasets to Understand Modern Android Malware Behaviours [INDUSTRY PROJECT]	Eirini S Anthi	Yulia Cherdantseva	2223-CMT400
Android malware has become a prevalent threat to mobile devices with the rise in banking malware and the increase in malware capabilities. Therefore, significant research has been done on Android malware detection, especially solutions that leverage machine learning. However, most of the training data from which these models learn contain outdated Android malware, making the models less effective at detecting recent unknown malware. This thesis covers the evolution and current threats of Android...[more]
Alhanouf Alsemairi	Creating Synthetic data for Stance Detection Tasks using Large Language models	Fernando Alva Manchego	Liam Turner	2223-CMT400
"Stance" refers to the viewpoint expressed in a piece of text towards a target. For example: Text: We live in a sad world when wanting equality makes you a troll #SemST Target: Feminist Movement Stance: Favor This task is usually tackled using supervised classification approaches. However, collecting and annotating datasets by humans can be time-consuming and expensive, thus may lead to a data scarcity problem. This project aims to use Large language models (LLMs) such as ChatGPT and Falcon as...[more]
Anil Singh	A system to compare and recommend IT consultants to Businesses to help them solve IT problems.	Stuart M Allen	Jing Wu	2223-CM3203
With many new start-ups and small to medium sized business being launched, it’s easier than ever to launch a business. But not without the challenge presented from online selling, tech problems, data storage issues etc. All these issues create a vast layer of complex IT problems for businesses to solve, many of whom are unequipped to. This leads to businesses turning to the web to source services from businesses and IT professionals to help solve their issue. But how do these companies know w...[more]
Aohan Li	Intelligent Fitness System: Integrating Advanced Action Recognition and Natural Language Processing for Personalized Guidance and Supervision	Sylwia Polberg	Víctor Gutiérrez Basulto	2223-CMT404
In today's society, with increasing awareness of health, more and more people are focusing on fitness exercises. Especially for beginners, scientific guidance and reasonable supervision are crucial due to a lack of fitness foundation. To meet this demand, we have developed an intelligent fitness system aimed at providing users with comprehensive fitness assistance services. The main functions of the intelligent fitness system are divided into two parts: human motion recognition and fitness train...[more]
Aryan Sajid Lakdawala	A Portfolio Builder Web Application	Catherine Teehan	Víctor Gutiérrez Basulto	2223-CMT404
The Portfolio Builder Web Application is an online platform that allows users to create and customise their digital portfolios. Users can create their portfolios for personal or professional purposes, and the web application will enable them to showcase their skills, experiences, and achievements. This web application will offer a user-friendly interface that enables users to add and remove entities and customise their portfolio according to their needs. The web application will offer a range of...[more]
Aryan Shetty	Identifying misinformation on Twitter using Machine Learning.	Federico Liberatore	Kirill Sidorov	2223-CM3203
Since the introduction of social media, misinformation or "fake news" has been a recurring problem in which posts would convey a message that is dubious in nature. This poses a great threat to our society as individuals could easily be swayed by the wrong information, leading them to believe things that are not real. To guarantee the information the user views is true, I would like to utilize machine learning algorithms and NLP (natural language processing) in python to determine and display pos...[more]
Billie O'Brien	Creating an application to incentivise young women into considering careers in STEM	Kathryn Jones	Yulia Cherdantseva	2223-CM3203
Having pursued an education within STEM it is very apparent that there is a huge lack of female participation in the field, particularly within Computer Science. There are many reasons for this including deeply ingrained traditional norms and stereotypes instilled from an early age, that cause women to be less supported in the field and fear participation. Finding a way to close this gender gap is essential to ensure diversity and equal opportunities for all genders in computer science and enab...[more]
Bingcong Yao	VR Stage Rehearsal Application	Dr Soumya Barathi	Dev Kant	2223-CMT404
A virtual reality application to help people alleviate stage fright and practise their presentation in front of a virtual audience. You will design and build this application to provide an effective platform to practise presentations while encouraging eye contact with the audience.[more]
Caleb Rowlands	CardiffUniMap: An Interactive Map for Cardiff University Students	Matthew Moloughney	Federico Liberatore	2223-CMT403
This project aims to begin exploring the potential for supporting the university experience of Cardiff University students by developing a prototype for an innovative interactive map web application. The current landscape of information at the university is fragmented and spread across multiple platforms, creating challenges for new students in navigating both the campus and available resources. This fragmentation also poses barriers to students' engagement with the academic community. Existing ...[more]
Calum Courtney	Deep Learning Guitar Transcription	Dave Marshall	Hiroyuki Kido	2223-CMT403
Many advanced sample libraries now exist that encompass a wide range of guitar types and playing styles. Other similarly exist for other musical instruments. The basic idea here is use a large range of sample libraries to deep learn playing style and instrument type. A deep learning approach to classify input as a playing style of guitar type Have learnt to identify such styles/type the information can be applied to input audio of an instrument to transcribe the varying articulations/styles...[more]
Calum McDowell	Art Generation	Paul L Rosin	Crispin Cooper	2223-CM3203
This project will attempt to create artwork in the style of some famous artists, so called generative art or computer art. There are many different methods possible as well as different artists to consider. The project could take as its basis, methods such as: Generation of Miro's Surrealism - Xiong & Zhang, 2016 Harold Cohen and AARON: Collaborations in the Last Six Years (2010–2016) of a Creative Life - Louise Sundararajan, 2021 Towards Layer-wise Image Vectorization - Ma, Zhou, Xu et a...[more]
Chaitanya Sanjay Taru	Leveraging Ensemble Learning for Calculating Expected Goals (xG)	Oktay Karakus	Xianfang Sun	2022Place-CMT400
The expected number of goals, or shortly the xG, is one of the most important football statistics that can be used to analyse today’s football in a quantitative manner. Its importance and usage have greatly increased during the last couple of years, and xG has a crucial meaning for football teams, scouting, betting firms, etc. Particularly, the xG is a football statistic which calculates the probability of a shot being converted into a goal. This allows match results to be explained better us...[more]
Chiwuike Anyiam-Osigwe	Mapping locations in texts	Chris B Jones	George Theodorakopoulos	2223-CM3203
The aim of these project is to produce a web application that allows the user to map the places mentioned in texts. The system will use named entity recognition (NER) and geocoding software such as that provided by SpaCy and the GeoPy libraries respectively to identify and geocode (generate the coordinates for) locations mentioned in a text provided by the user. The locations will then be plotted on a map (using a background map such as that of OpenStreetMap).[more]
Chloe Andronicou	Fashion Sustainability Application using Augmented Reality	Dr Katarzyna Stawarz	Hantao Liu	2223-CMT400
The environment today faces many threats causing its degradation and the fashion industry is one of those contributing threats. This project aims to address some of the issues by facilitating sustainable fashion efforts and raising awareness through Augmented Reality (AR) technology. A web application is developed that decreases excessive unnecessary clothing purchasing by having users create a digital inventory of their clothing. The application aims to educate the user regarding sustainability...[more]
Chun Ting Justin Lo	AI-powered Sign Language Application - Signify	Fernando Alva Manchego	Steven Schockaert	2223-CMT403
Proposed App Name: - Signify Application Aims: - The sign language application allows users to easily create and manage their own custom sign language translation models using AI. Most existing sign language applications that are mainly intended for translation and educational purposes only support widely recognized sign languages like ASL, leaving many highly localized and specialized sign language terms unsupported. Such as sign language terms for a local community and for the electrical engi...[more]
Craig Thompson	Smartwatch for Indoor Navigation	Charith Perera	Matthew J W Morgan	2223-CM3203
Navigating large and complex buildings can be challenging for individuals with communication difficulties, particularly in unfamiliar environments like university buildings or museums where seeking assistance may not be an option. This report covers a 5 month-long project to assist with this problem by creating an indoor navigation device using a combination of the open-smartwatch and many BLE (Bluetooth Low Energy) Beacons. The project was implemented in the University's ABACWS building with be...[more]
Dan Grimsey	Using Minecraft Education Edition to Teach the Uses of Radiation	Catherine Teehan	Neetesh Saxena	2223-CM3203
This report presents the development and implementation of a Minecraft Education Edition world designed to teach students about the dangers and applications of radiation, aligning with the new 2022 Welsh curriculum for Science and Technology. The project is aimed to create an engaging and effective learning experience by leveraging Minecraft’s immersive nature. The report outlines the objectives of the project, including gaining knowledge about teaching in conjunction with the new curriculum, ...[more]
Daniel Dominguez	Using multi-modal AI to predict water quality	Oktay Karakus	Steven Schockaert	2223-CMT403
Explore the use of training machine learning algorithms using multi-modal datasets through the use of ecoregion categorizations and type of body of water for measuring water quality parameters. Explore the use of using more remote sensing indexes than what is available in Landsat base images.[more]
Darcie Lowe	Knowledge Management via Natural Language Processing: Document Discovery in the Wales Safeguarding Repository	Alun D Preece	Jianhua Shao	2223-CM3203
The Wales Safeguarding Repository (WSR) is an online collection of safeguarding review reports containing valuable information about learning experiences in tackling serious incidents involving children and vulnerable adults. Research in the School of Social Sciences at Cardiff University has highlighted the need for better organisational learning to emerge from these reviews, in order to enhance the future safeguarding practices of professionals such as police officers, social workers and those...[more]
David Diaz Clifton	Developing and application to aid users overcome their IT distractions and addictions and improve productivity	Nervo Verdezoto Dias	Yipeng Qin	2223-CM3203
For my Design thinking and prototyping for UX module, I aim to prototype an application for users that use technology in their work. I hope this app makes them aware of how their addictions affect them when working, and pushes them to plan out their tasks and stick with them. There are a couple of existing technologies that help with users productivity such as Notion, notes and stickies and calendars to help set tasks within timeframes, and there are also apps that fully block apps on users ph...[more]
Dayang Chen	Collaborative Work Web Site for Research Groups	Alia I Abdelmoty	Stuart M Allen	2223-CM3203
This project will design and build a web site to allow research groups to work collaboratively together. A user-centred approach will be adopted to define the requirements and design and evaluate the project. Beyond Web page design, the site needs to allow researchers to contribute information about their research projects, work together on a shared reference store, keep track and provide alerts to relevant new publicatins/journals/calls. This is also an exercise in HCI and evaluating the usabi...[more]
Devajith Valaparambil Sreeramaswamy	Mesh Filtering Acceleration with GPUs: A Comprehensive Analysis	Bailin Deng	Frank C Langbein	2022Place-CMT403
Mesh filtering is an important tool for cleaning up mesh models obtained from 3D scanning. Its main goal is to remove noises from the mesh shape, while preserving geometric features such as sharp edges. Various techniques have been proposed to perform mesh filtering on CPU. State-of-the-art methods can produce robust results but may take a long time to complete the task on large meshes. The goal of this project is to modify an existing mesh filtering method to run them on GPU instead of CPU, in ...[more]
Emin Safa Tok	Motion Planning for Robot Object Manipulation Using Machine Learning	Juan Hernandez Vega	Padraig Corcoran	2223-CMT400
Motion planning is a relevant research area in robotics and its purpose is to find a collision-free path from a start to a goal robot configuration. Especially for manipulator arms, planning problems require contact-rich interactions between a robot and the objects in the environment without collision. There are several techniques to solve these problems, such as search-based and sampling-based algorithms. Besides these techniques, learning-based methods are used to enhance efficiency at solving...[more]
Emma Jenkins	Developing ML model to detect drive-by download on Twitter and uncovering cyber criminals tactics.	Amir Javed	Dave Marshall	2223-CM3203
Stage 1 Data collection & Experimental setup You will need to identify 2-3 popular events such as Covid or trending topic to collect data from Twitter or other social platforms like Instagram. I'll let you decide on the platform. You will need to create a setup to store the data. Preprocess the data for the next stage.- identify disinformation Create an account with Virus total, multiple accounts so you can check many URL’s or you can use Cuckoo honeypot to analyse the URL [this is relat...[more]
Enrique Paulino	Active glasses	Kirill Sidorov	Xianfang Sun	2223-CM3203
This project aims to develop an algorithm for real-time video enhancement using a camera to simulate a pair of “virtual active glasses”. The goal of this system is to enhance the visual experience for the user through different computer vision techniques in order to bring the object of focus to the forefront. Techniques could include blurring to create a bokeh effect and increasing the brightness, contrast, or sharpness of the object of interest. Other techniques to bring the object of focus...[more]
Ethan Le Sage	Predicting the Winners of Popularity Contests using Twitter Sentiment Analysis	Fernando Alva Manchego	Natasha Edwards	2223-CM3203
Twitter sentiment analysis for I'm a Celeb to predict the winners before, or one the day of, the finale. I already made something over the summer which succesfully predicted the winners of Love Island four weeks before the finish (while the odds were 3:1). Could be adapted to be far more effective, and potentially work for any audience-voting game show. Unlimited possibility for increased complexity and features. Coded in Python with sentiment analysis models, machine learning models, and more...[more]
Feiran Su	Connecting Cultures Through Food: A Recipe Sharing Platform for Cardiff University Students	Martin Caminada	Richard Booth	2223-CMT403
Introduction: Cardiff University is home to over 7,530 international students from over 138 countries. These students may find it difficult to adapt to local food when studying abroad, resulting in a sense of isolation and disconnection from their home culture. To address this, it is proposed to create a recipe sharing platform that enables students to share and learn traditional dishes from their home country. Problem Statement: The lack of access to home-cooked meals can lead to a sense of is...[more]
Fin Bignold-Jordan	The Community Knapsack: Exploring Optimisation Algorithms For Combinatorial Participatory Budgeting	Richard Booth	Natasha Edwards	2223-CM3203
Participatory budgeting is a democratic approach to decision-making in which citizens or residents of municipalities and communities directly vote on the allocation of public resources to projects and proposals. The combinatorial model of participatory budgeting, in which projects must either be fully or not at all funded, is widely solved using an arbitrarily bad greedy approach which fails to utilise the resources maximally. This project aims to improve participatory democracy by investigating...[more]
Fin Cottle	Evaluating and Extending a Repurposed Natural Language Machine Learning Model for Symbolic Music Completion	Dave Marshall	Yipeng Qin	2223-CM3203
Natural Language Processing (NLP) is a rapidly growing machine learning technology used across hundreds of industries. One of the main techniques within this field is BERT (Bi-Directional Encoder Representations of Transformers), a deep learning model designed to handle tasks such as question answering, sentiment analysis, text summarisation and more. Applying these techniques within music instead of a commonly spoken language is a relatively new concept with the potential to revolutionise the c...[more]
Finn Trowell	MelAi: LibMelee AI via Deep Reinforcement Learning	Frank C Langbein	Sylwia Polberg	2223-CM3203
Super Smash Bros. Melee is a platform fighting game in which two or more players can fight on 29 unique stages with 26 different characters. It has become an ever-growing competitive fighting game, fostering an Esports scene, thanks to its uniquely precise analog controls. These controls allow players to perform frame-perfect actions at blistering speeds, perfect for artificial intelligence to exploit. The goal of this project is to create an artificially intelligent agent to play Super Smash ...[more]
Fred Cozzi	Autonomous Robot Development	Eirini S Anthi	Yazmin Ibanez Garcia	2223-CM3203
The aim of the project is to create a floating robot such as (https://www.tate.org.uk/whats-on/tate-modern/hyundai-commission-anicka-yi) which is able to move autonomously and perhaps even interact with people. The student will be given access to the equipment needed including sensors, 3D printer, etc.[more]
Gaia Anna Rosaria Lo Giudice	A Comparative Sentiment Analysis of Climate Change Discourse in the United Kingdom and Italy	Alun D Preece	Nervo Verdezoto Dias	2022Place-CMT403
The damage misinformation and conspiracy theories can cause has risen to prominence in recent years. During the COVID-19 pandemic, conspiracy theories spread via social media have led to numerous disruptive events including attacks on 5G towers and direct action by anti-vaccination groups. In the political arena, QAnon conspiracy theorists were involved in the January 2021 attack on the United States Capitol, threatening the legitimate democratic process of a major nation state. In 2019, the Wo...[more]
Guanfei Wang	Image recognition to augment models of sustainable transport	Crispin Cooper	Irena Spasic	2223-CMT403
Here in Cardiff we have developed simulations of pedestrian and cyclist behaviour that are used for sustainable transport modelling worldwide (www.cardiff.ac.uk/sdna) however these efforts are often limited by the data available. This project is about harvesting information from crowdsourced street images, and automatically detecting features such as cycle lanes and pavements to better inform transport simulations.[more]
Hang So	Implementation of a data privacy protection tool for relational data	Jianhua Shao	Shancang Li	2223-CM3203
As an increased amount of data being gathered and stored, how to protect the private information contained within such data sets becomes an important issue. One of the recent approach to addressing this issue is called k-anonymisation, which attempts to make any record in a data set identical to at least k-1 other records (hence no individual could be identified). This project aims to implement a software tool based one of the exisiting k-anonymisation algorithms, to help anonymize relational da...[more]
Haohan Wang	Learning to Detect and Recognise Road Signs	Yukun Lai	Federico Liberatore	2223-CMT400
To improve driving experience, computer vision techniques are widely used for providing assistance. The project aims to develop a prototype that implements some machine learning technique to detect and recognise road signs from real world camera footage. Such technology is also an essential component for autonomous driving.[more]
Hebe Wrench	Grasping Regions through Learning and Computer Vision Techniques	Juan Hernandez Vega	Stuart M Allen	2223-CM3203
Grasping objects is one of the basic and most common tasks that a robot (such as manipulator arms or mobile manipulators) has to do. While it is considered a simple task for humans, sometimes it could be a challenging robot operation. For a robot, grasping an object normally involves perception to detect the object, planning to approach and grasp the object, and control to drive the robot through the planned motion. With the increasing popularity of learning techniques, new frameworks/libraries ...[more]
Hugo Huang	Doom-playing AI via Deep Reinforcement Learning	Frank C Langbein	Fernando Alva Manchego	2223-CM3203
Training artificial intelligence (AI) agents to play games directly from high-dimensional sensory inputs (like visuals or audio) was widely considered as one of the greatest challenges of reinforcement learning (RL) until Deep Q-Network (DQN) was proposed. Even today, training agents to play first-person shooter (FPS) games like Doom is a non-trivial task and in this project I will attempt to create an AI agent that plays Doom via a Deep Q-Learning-related approach. The goal is to utilize the Vi...[more]
Ioannis-Marios Stavropoulos	Robot Navigation Commands via Augmented Reality Interfaces	Juan Hernandez Vega	George Theodorakopoulos	2223-CM3203
New advances in sensors and algorithms are allowing robots to navigate different types of environments (e.g., underwater, space, among others). To do so, robots generally receive a destination (goal), and then they plan collision-free paths to reach such a destination. An important aspect when commanding a robot to conduct such navigation tasks is the modality used to command the robot, i.e., the mechanism that a person employs to give a robot the destination. The main goal of this project is ...[more]
Isaac Powell	Intelligent game-playing opponent	Andrew Jones	Sylwia Polberg	2223-CM3203
This project involves implementing an intelligent opponent for some game of skill such as Chess, but the game chosen would preferably be an unusual one. You may use whichever programming language you prefer (Java or Python would be suitable choices, or you might wish to use something like Prolog, which would lend itself very well to a task such as this). You may like to implement generic software components that can be used in more than one game. You will be expected to pay particular attention...[more]
Ishana Raina Ghosh	Assessing the Impact of Contextual and Entity-Specific Features in a Spoiler Prediction Model	Hiroyuki Kido	Alia I Abdelmoty	2223Place-CMT404
This dissertation presents an in-depth analysis of how the integration of BERT embeddings and Named Entity Recognition (NER) features in a deep learning framework influence the performance of a machine learning model in a spoiler detection task given the generalised hypotheses of their robust performance and high accuracy in complex tasks related to Natural Language Processing (NLP). The primary aim is to dissect and quantify the contributions of these two pivotal elements—contextual language ...[more]
Jack Boyer	Explainable Malware Detection based on Machine Learning Algorithms	Tingting Li	Amir Javed	2223-CM3203
Many cybersecurity challenges now can be dealt with by using the power of AI, from automatic identification of malware to intelligent defence systems combating ever-changing cyberattacks. However, many cybersecurity companies find it difficult to trust AI solutions and leave protection in the hands of a mysterious black box. Therefore, this project aims to develop AI-enhanced cybersecurity solutions which are accountable and transparent. For instance, a framework to provide an explanation to mal...[more]
Jackman Brookes	Sample based Bass Guitar Transcription	Dave Marshall	Charith Perera	2223-CM3203
Transcriptions are the written form of music and are essential for the performance, study, and preservation of music over the years. Traditionally, the process of transcription is performed by a skilled musician who listens to a piece of music and transcribes what notes are played. For most people this takes years of training their ears to recognise notes but with the advent of automated transcription algorithms, transcribing music is more accessible than ever before. Current leading transcripti...[more]
Jack Sanderson	Calculating the Impact of EIP-1559 on Transaction Fee Variance	Yuhua Li	Natasha Edwards	2223-CM3203
Every blockchain has a "fee market"; this market determines the cost per byte of block space. In the past, during both the 2017 and 2021 bull markets, many blockchains used an antiquated fee market that featured a simple auction system. This caused enormous fluctuations in fees greatly harming the user experience and even causing issues with applications built on top of the network. On August 5th of 2021, the Ethereum blockchain hard-forked onto a version that featured a more advanced fee market...[more]
Jack Smith	Compare the efficacy of methods for estimating the CO2e of a transaction using data available via a banking API.	Crispin Cooper	Irena Spasic	2223-CMT403
The goal: predict the carbon footprint of a transaction using the information collected by a bank, available via an API (Starling or Monzo). The data: transaction amount, vendor category (groceries, travel, eating out, etc.), vendor name. Potential methods: 1. Calculate an average CO2e/£ for each category and apply it to the amount spent. 2. A one-off user-completed questionnaire which informs future estimates. 3. Use the vendor name to find out what they sell (scrape data? ask ChatGPT?) and u...[more]
Jake McNeill	Developing an absolute beginner friendly Capture The Flag platform	Eirini S Anthi	Padraig Corcoran	2223-CM3203
This project will aim to develop a CTF platform for absolute beginners. A key requirement of this project will be an analysis and categorization of a range of different challenges by difficulty and techniques. The platform will also collate and include relevant sources to users to help them understand how to approach solving such challenges.[more]
James McLaughlin	Music makes you run faster...	Martin J Chorley	Theodoros Spyridopoulos	2223-CM3203
Last.FM is a music service which allows you to access a list of a user's recently played tracks. Multiple services exist for tracking exercise (Strava, Fitbit, Runkeeper, ...) and associated data (heartrate, pace, ...). I'd like to create an app that allows you to compare your exercise data with your recently played music over a large time period to determine whether the music you were listening to has any impact on your performance in the exercise. This could be created as a secure web applic...[more]
James Price	Collaborative Work Web Site for Research Groups	Alia I Abdelmoty	Amir Javed	2223-CM3203
This project will design and build a web site to allow research groups to work collaboratively together. A user-centred approach will be adopted to define the requirements and design and evaluate the project. Beyond Web page design, the site needs to allow researchers to contribute information about their research projects, work together on a shared reference store, keep track and provide alerts to relevant new publicatins/journals/calls. This is also an exercise in HCI and evaluating the usabi...[more]
James Tapp	[Industry] Anatomy highlighting in lung ultrasound Images via deep learning	Oktay Karakus	James Osborne	2223-CMT403
(DSA-23) Ultrasound is a non-invasive and safe (non-ionising) imaging medium which may be used to diagnose and monitor many conditions, including the assessment of lung pathologies. At Intelligent Ultrasound we enhance the effectiveness of Ultrasound-based diagnoses through the use of AI and in this project, we’d like your help! We have a large (though only partially labelled) dataset of lung ultrasound images, containing tens of thousands of images obtained from hundreds of patients. The p...[more]
Jamie Groom	Predicting Expected Career Rating (xCR) of NCAA Players after NBA Drafts	Oktay Karakus	Bailin Deng	2223-CM3203
This project aims to create a historical data set for NCAA player statistics and combine this with NBA draft information to predict how a rookie can do in challenging NBA? A "Boom" or a "Bust"? The created data set will be used to train a machine learning algorithm and the predicted target will be a "rating point" - namely the expected Career Rating (xCR) - that potentially shows how this player is going to climb the steps in the early years of their career. This project requires handling a hu...[more]
Jayant Misra	Robust Quantum Control via Optimising a Stochastic Target Functional	Frank C Langbein	Matthew J W Morgan	2223-CM3203
An important part to turn quantum physics into technologies is the robustness of the quantum process, e.g. for quantum computing, sensing or simulation. Quantum control provides the means to steer quantum processes, but often it is focused on high fidelity and not also robustness of the process. We developed a stochastic measure to judge the fidelity and robustness of a quantum process. The aim of this project is to investigate methods that optimise this measure directly to find a robust, high-f...[more]
Jiahao Li	Image super resolution	Xianfang Sun	Padraig Corcoran	2223-CMT400
Image Super-Resolution (SR) is a fundamental class of image processing techniques in computer vision to recover a HR image from the LR one, which improves visual per- ception and enhance details of the image. Deep-learning-based methods have shown impressive performance in SR tasks. However, the most previous SISR methods only treat SR of different scale factors as independent tasks. They train a specific model for each scale factor which is inefficient in computing. This work proposes the Hybri...[more]
Jiaqi Wang	Image decolorisation	Paul L Rosin	Carolina Fuentes Toro	2223-CMT400
If a colour image is converted to gray scale then significant features are sometimes removed (i.e. two coloured objects with the same brightness will look the same in gray scale). This project will investigate approaches for overcoming this. See Gooch et al.'s "color2gray" system on the web for more details. In addition, if time permits, a similar approach will by applied to modify the hue in colour images, which should generate artistic images in a "Fauvist" style. Some other related papers: ...[more]
Jingtao Yu	Generating Images from Audios Using AI	Bailin Deng	Usashi Chatterjee	2223-CM3203
Given an audio clip, can you draw the scene corresponding to the audio? This project aims to solve this problem using deep learning techniques. We will first use audio captioning models (for example, see https://github.com/audio-captioning/audio-captioning-papers for a list) to generate a free text description for the audio clip. Then an image can be generated from the text description using text-to-image generation models (see https://stability.ai/blog/stable-diffusion-public-release for an exa...[more]
Joel Harris	Using Machine Learning to determine the likelihood of a shot resulting in a goal in football.	Hiroyuki Kido	Yazmin Ibanez Garcia	2223-CM3203
The term expected goals (xG) is a statistical measurement in football to help determine the likelihood of a shot resulting in a goal. An xG model uses historical data of shots to estimate the likelihood of a goal on a scale between 0 and 1. The xG is a crucial measurement in the sport as it is the most accurate predictor of future team and player performance that is available. Expected goals has previously been calculated using equation modeling. However, more accurate results can now be determ...[more]
Joe Liu	PECS Appliaction for Children with Autism	Fernando Loizides	Stuart M Allen	2223-CM3203
This app will be used in the real world. You will have access to not only your supervisor but also to the client for weekly meetings. Currently, therapists communicate with some children through pen, paper, and pictures that are printed and posted on a board. This communication is basic, like "I want to eat an apple". We would like to create a digital version of this. We need a web app built that will include rich interaction like taking pictures of common items that a child can understand and...[more]
John Baker	Designing a comprehensive and bespoke UX focused solution that enables the use of a martial arts syllabus and grading system in a low bandwidth network.	Kathryn Jones	Stuart M Allen	2223-CM3203
This project will design, develop and evaluate a comprehensive and bespoke user experience focused system containing a martial arts syllabus that is accessible in low bandwidth network areas. For example, when at a dojo or a gym, usually the throughput isn’t transferring enough data fast enough to support looking up patterns or techniques. In addition to this, it will support an intuitively designed system for the syllabus and grading aspects, regardless of bandwidth speeds, which will help pr...[more]
Jonathan Ayooluwa Ayodele	Understanding The Impact of Risk Perception on Cybersecurity Training Effectiveness in Small and Medium Enterprises (SMEs)	Amir Javed	Yulia Cherdantseva	2223-CMT400
The aim of this project is to understand people's risk perception, especially in small and medium enterprises (SMEs), influence their susceptibility to awareness training and implementation of other security measures ============================================= How do people's risk perception, especially in small and medium enterprises (SMEs), influence their susceptibility to awareness training and implementation of other security measures? This should include consideration of scenarios where ...[more]
Jonny Streatfeild-James	Cyber Security Education Online Portal	Yulia Cherdantseva	Fernando Alva Manchego	2223-CM3203
Cybersecurity is an increasingly important factor of everyday lives, with more users gaining access to the internet and more risks becoming easier to fall victim to. Moreover, the jobs required in the cybersecurity sector are not being filled due to the lack of students being educated about cybersecurity and the lack of professionals qualified to teach the students. This project aims to address these issues by creating a secure website to introduce and educate children on cybersecurity. The webs...[more]
Joseph Llewelyn	Use of AI and image processing to assist with eye medicine compliance	Shancang Li	Paul L Rosin	2223-CMT403
Monitoring the use of eye drop medication is key to the management of Glaucoma patients. We have developed a system for gathering video images of the administration of eye drops and would like to develop an AI tool to detect the correct delivery of eye drops. This Project will be supervised by Richard Perks, School of Engineering.[more]
Joshua Smith	Development of a Gamified Cybersecurity Learning Portal for Key Stage 3 Students	Yulia Cherdantseva	Xianfang Sun	2223-CM3203
Cybersecurity is a growing concern in the UK due to the rising threat of cybercrimes, yet there is a lack of individuals who possess the necessary skills to mitigate them. The failure to recruit young people into cybersecurity career pathways has been identified as a reason for this problem. This project aims to address the disconnect between the increasing demand for cybersecurity professionals and the limited size of the future cybersecurity workforce. To encourage more young people to pursue ...[more]
Juliette Radley	Researching Socially Responsible Human-Robot Interactions for Use in Companionship with Older Adults	Carolina Fuentes Toro	Federico Liberatore	2022Place-CMT404
This MSc dissertation aims to research how to design socially responsible Human-Robot interactions for use in companionships with older adults. To comprehensively achieve this objective, interviews with older adults and their carers were conducted and user profiles, storyboards, design requirements, and a prototyped interaction were created. The main research method used to acquire primary qualitative data was semi-structured interviews. These interviews were supplemented by further discussions...[more]
Junhao Zhou	Age estimation Based on Deep Learning approach	Frank C Langbein	Hiroyuki Kido	2223-CMT403
Age detection based on face recognition is a technology that uses computer algorithms and machine learning techniques to analyze facial features and predict a person's age. The system typically analyzes various facial characteristics such as wrinkles, skin texture, and facial contours to estimate the age of an individual. This technology can be used in a variety of applications, such as marketing and advertising, security, and healthcare. It has the potential to improve personalized services, op...[more]
Junlin Li	Improved YOLOv5 with attention mechanism research on object detection without forms	Charith Perera	Alia I Abdelmoty	2223-CMT404
Deep learning methods have been shown to outperform prior machine learning techniques in multiple domains, including cases that arise in computer vision, such as convolutional neural networks. Object detection has always been a hot area in computer vision. The YOLO algorithm is based on a 2015 CVPR article by Joseph Redmon called You Only Look Once: Unified, Real-Time Object Detection.YOLO is a more advanced algorithm based on convolutional neural network, which is better than the traditional ...[more]
Kacper Bulynko	Deep Learning Guitar Tunings	Dave Marshall	Omer F Rana	2223-CM3203
Guitars can be tuned to a a variety of tunings. This is especially common in certain genres of music, notably folk music and heavy metal, for example. The basic idea here is to take recordings on music played in known guitar tunings. I have database of known songs in know tunings by a variety of artists and have the associated recordings. A deep learning system can then be trained with input samples (many short snippets from every recording) and trained to classify the tunings. New music can t...[more]
Khai Pong Teoh	Development of an effective Automatic Number Plate Recognition using YOLO detection	Bailin Deng	Liam Turner	2223Place-CMT403
Automatic Number Plate Recognition (ANPR) is a technology that employs optical character recognition to scan, detect, and read vehicle registration plates. By converting images (captured by cameras) of vehicle plates into machine-readable text, ANPR plays a crucial role in various applications, from traffic monitoring to security enforcement. Despite significant advancements in computer vision through deep learning, many solutions remain insufficiently robust for real-world applications. To addr...[more]
Kieran Williams	Carbon Reducing Service Provider Marketplace on the Blockchain	Omer F Rana	Oktay Karakus	202223-CM3203-RP
The project is centred around developing a unique website-based marketplace platform that acts as a conduit between landlords and service providers. The goal is to provide an ecosystem where eco-friendly services (jobs that reduce carbon emissions) are easily accessible, promoting a reduction in the carbon footprint of properties. The marketplace platform is not only a means of connection but also a tool for endorsing sustainable practices in the property sector.[more]
Kiya Bhayani	Algorithms & Echo Chambers: Can Twitter Predict an Election?	Stuart M Allen	Irena Spasic	2223-CM3203
Twitter has transformed the way individuals communicate and share information. Politics is one area where it has had a significant influence, becoming a tool for political campaigns, political analysts, and voters to express themselves, mobilise supporters, and obtain real-time information. This study aims to look at the relationship between Twitter sentiment and election polling data, specifically whether Twitter sentiment can be utilised to forecast election outcomes accurately. The study imp...[more]
Kyriakos Nicolaou	Unity Game [Industry]	Fernando Loizides	Bailin Deng	2223-CM3203
We are working with Google to create a simple unity game that takes inputs from a natural language model and depending on the accuracy of the word makes a character move (jump how high depending on the accuracy level). The student would take that input and create the game in unity.[more]
Lawrence Walker	Using Object Detection to Identify Human Movement	Charith Perera	Oktay Karakus	2223-CM3203
The idea of my project is to test the capabilities of small computing power. I will use a Coral Development Board, connected to three Coral Cameras, to identify objects in a small lab environment. The Coral Development board will contain datasets that I have researched to help it identify a human or object that moves past a camera. The cameras will be placed in three different areas in the same lab, so that each camera's ability can be tested on how well it detects a human in the picture. The go...[more]
Lewis Carter	A driving Tool to help drivers with the rising cost of Fuel with a In-depth fuel tracking and statistics web application	Padraig Corcoran	Paul L Rosin	2223-CM3203
In this project I will create a web application which helps drivers track their fuel. It will give feedback on their MPG (Miles Per Gallon), MP£(Miles per Pound) and give a driving efficiency percentage, this is a few of the possible statisitics that can help the average driver cut down on the amount of fuel they are using. This has economical and environmental benefits as this will help driver use less fuel, spend less money and emit less emissions. Another way i see this project progressing i...[more]
Lewis Richards	A Video Game to Raise Cybersecurity Awareness - Digital Detective	Eirini S Anthi	Stuart M Allen	2223-CM3203
General IoT users lack the knowledge and awareness of Cyber security to adequately defend themselves from various attacks, for example, viruses and ransomware. This project aims to provide an interactive learning environment, in the form of a game, about various aspects of Cyber security, for example how to identify phishing attacks. The importance of this project is to aid in spreading cyber security awareness across general IoT, in an entertaining and interactive manor.[more]
Liban Ahmed	Evaluate the Effectiveness of “off-the-shelf” Intrusion Detection and Prevention Systems in protecting Critical National Infrastructure against common cyberatta	Andrew Hood	Amir Javed	2223-CMT400
I have created a test bed network of virtual machines consisting of an attacker VM loaded with my attacking scripts, a poll vm which mimics a key ICS control station through using Qmodbus to send instruction to the slave and a Slave VM which contains ModbusPal and mimics an ICS environment with four coils to turn on/off functionality. The two trusted devices on the network here are the Poll and Slave VM. The Attacker Kali acts as an network intruder and can send attacks directly to the poll and...[more]
Lihao Liao	Cyber Security Education Online Portal	Yulia Cherdantseva	Yuhua Li	2223-CM3203
The aim of this project is to develop a web-application for hosting cyber security educational material for audiences of different ages starting from primary school and up to the MSc level, and beyond. The application will have must allow access to educational material, allow filtering of material based on different criteria. User roles in the application will include: Admin, Educator, and Participant. The access permissions will vary based on the role. The material stored will vary by type and ...[more]
Luc Woodman	Fingerprinting The Networks of Users Disseminating Malware on Twitter	Amir Javed	Padraig Corcoran	2223-CM3203
Abstract Online Social Networks (OSNs), such as Twitter, have become the become the breeding grounds for the spread of malicious content, particularly through the dissemination of embedded URLs. This research paper aims to investigate the characteristics of malicious users involved in spreading malware on Twitter and compare their behaviour to emotionally charged events. Scope Answering the following three research questions. -Is there a presence of organised behaviour distributing malicio...[more]
Luke Clifton	A Piano Learning Companion App	Andrew Jones	Fernando Loizides	2223-CM3203
The application would generate a random sequence of chords or notes. This would be conveyed to the user in their desired method(s); the notes represented as letters, the musical sheet representation and/or the sound being played for example. The user is then expected to play back the sequence, and using the device's microphone the application can verify if the user has done so correctly. Repetition of this exercise can help solidify a user's learning, and would help them learn to read sheet musi...[more]
Luke McCarney	Designing an Interactive Application to Promote Healthy Eating Practices at Home	Nervo Verdezoto Dias	Hiroyuki Kido	2223-CM3203
Prerequisites: -Basic knowledge of Human-Computer Interaction and user-centered and/or participatory design methods -Experience developing mobile and/or web applications in several platforms (e.g., Android, HTML5 & Javascript, etc.) -Desirable experience with Spanish (not mandatory) Digital health technologies are becoming increasingly prevalent to support self-care practices aiming to enhance people’s health and wellbeing. However, their potential to promote healthy eating practices amon...[more]
Marcus Anderson	Volleyball Statistics Tracking Application	Jing Wu	Matthew J W Morgan	2223-CM3203
Every volleyball team should have a qualified coach to help them play however not all do. Some volleyball clubs this year lacks an experienced, qualified coach to help us succeed in the leagues we play in; this stat tracking app is made to allow teams to players to analyse their play and make decisions in real time even without a proper coach, allowing for better decisive plays and a clearer view of what areas of interest players may need to work on (e.g. Passing, Spiking, Setting). My motivat...[more]
Mariia Semenenko	Low Vision Webpage Layout Modifier: A Web Extension for People with Visual Impairments	Paul Goddard	Alun D Preece	2223-CMT400
This project is dedicated to advancing web accessibility for individuals with low vision by creating a web extension that addresses prevalent challenges and incorporates recent literature recommendations. The primary objective is to establish a universally accessible webpage layout, filling the existing gap between current low-vision aids and real-world challenges. This study aims to enhance the consistency of web design and development, ultimately contributing to increased user confidence among...[more]
Matilde Pinto Vicente	Role of Social Robots in Everyday Support and Companionship for Older Adults	Carolina Fuentes Toro	Juan Hernandez Vega	2223-CMT403
Currently, the development of robots for supporting older adults focuses on health care, medicine reminders, coaching for physical activity, and social interaction among others. It is expected that robots would support older adults to live independently for longer and many opportunities for supporting small everyday tasks emerge. This project aims to explore how a robot can support older adults with everyday tasks, and in particular cooking tasks. Based on a review of existing literature, the pr...[more]
Matthew Battagel	Malware Detection within Object Storage	Theodoros Spyridopoulos	Matthew J W Morgan	2223-CM3203
The rapid increase in data production has amplified the necessity for robust and secure data storage systems. MinIO, an open-source object storage platform, offers cost-effective and scalable solutions but lacks an integrated malware detection feature, thereby posing potential risks to user data. This project seeks to augment MinIO by incorporating a scalable malware detection system, with the goal of maintaining MinIO's performance and scalability. In essence, this project aspires ...[more]
Matthew Robinson	Designing a Mobile Application to Promote Better Sleep Habits Using NFC-Assisted Routines	Dr Katarzyna Stawarz	Catherine Teehan	2022Place-CMT403
With an increasing prevalence of sleep disorders and the continued impact of mobile phone overuse on sleep quality, there is a need for technological interventions that effectively promote sleep health, whilst understanding the impact the device itself has on sleep hygiene. This study aimed to implement a solution that balances these two ideas whilst implementing a novel approach of Near-Field Communication (NFC) assisted sleep routines.[more]
Max Christopher Christopher	Time Series Data Generation/Simulation- Home sensors example	Alia I Abdelmoty	Juan Hernandez Vega	2223-CM3203
This project will address the problem of limited realistic data availability for testing and evaluation in the domain of home sensor data. The collection of data in a home setting is limited as it is seen as intrusive and a privacy risk. There is, however, the need to examine and analyse large data sets, particularly for data mining and ML tasks. This project will consider some data collected over a period of almost a a year from two homes equipped with different environmental sensors and will d...[more]
Max Green	Hearing the future: Predicting the next piece of audio	Dave Marshall	Bailin Deng	2223-CM3203
The basic ideas is based on a sequence of audio can you predict the next few seconds. Deep learning networks (E.g LTSM, Recurrent Neural networks) can be used. Training data is abundant: any audio of a few seconds and be utilised. Take a few second segment and use this build a model that predicts the next segment. A variety of interesting questions need to researched: * What the the best format for the input audio * What type and configuration of network is best * Format of training data: how ...[more]
Melanie Meijer	Automated Cyber Defence by Deep Reinforcement Learning	Tingting Li	Ian M Cooper	2223-CM3203
Defending cyber security is a significantly unfair game between defenders and attackers, as defenders need to be cautious all the time to detect and react to every single attack, whilst attackers only need to strike once at any time. Rapid development in Artificial Intelligence (AI) provides the potential for distributed, adaptive defensive measures at machine speed and scale. It is possible now the defender can be trained as an intelligent agent to develop strategies to effectively respond to a...[more]
Mihail Bors	Anatomical Segmentation of Prostate MR Images	Frank C Langbein	Jing Wu	2223-CM3203
The aim of this project is to investigate approaches towards automatically segmenting the prostate in MRI datasets, based on the PROMISE12 grand challenge and our own, internal data set (other data sets are also available). This is part of a research project for early-stage prostate cancer detection. You may develop your own approach or test and then extent already published approaches. Which technique you are using is your choice, but it is likely that a deep learning approach (e.g. U-NET and v...[more]
Mika Davis Jupp	A smart speaker tool to 'beckon' people away from their desks	Stuart M Allen	Juan Hernandez Vega	2223-CM3203
Spending too much time at our desks is bad for us. There are lots of tools to encourage people to get up and move around, but, as these are often computer-based, they are not always successful. In this project, you will develop an alternative system. It will use smart speakers located around people's homes to encourage them away from their desks. This client-server system will allow people to set up a schedule on a website and have these preferences reflected in the way that their smart speaker ...[more]
Moath Majdi M Khawaji	Risks Assessment and Advance Alert Notifications for Smart Grid	Neetesh Saxena	Yulia Cherdantseva	2223-CMT400
This project aims to understand a smart industrial system (e.g. smart grid) use case scenario, predict risks for critical assets, perform modelling assessment, and develop an advance alert notifications tool for situational awareness.[more]
Mohammad Parvareshnia	Music makes you run faster...	Martin J Chorley	Matthew J W Morgan	2223-CM3203
Last.FM is a music service which allows you to access a list of a user's recently played tracks. Multiple services exist for tracking exercise (Strava, Fitbit, Runkeeper, ...) and associated data (heartrate, pace, ...). I'd like to create an app that allows you to compare your exercise data with your recently played music over a large time period to determine whether the music you were listening to has any impact on your performance in the exercise. This could be created as a secure web applic...[more]
Mohammed Fayaz Ansar Jelani	Aligning generative transformer models with fine-tuning and supervision	Yuhua Li	Yazmin Ibanez Garcia	2022Place-CMT400
This research project centres on the issues and possible solutions related to the utilization of generative transformer models (GTM) for creating generative chatbots. Although GTMs offer dynamic and engaging conversations, they also present challenges in terms of controlling responses and the risk of malicious attacks. The objective of this research is to align GTMs with core business applications and prevent deviations from the intended functionality. The research goals encompass fine-tuning GT...[more]
Mohammed Shaad Mehboob Matcheswala	Creating an Adaptive Defense Architecture using an Adaptive Honeypot Algorithm and Network Traffic Classifier	Amir Javed	Andrew Hood	2022Place-CMT400
In this age of digital transformation, more and more businesses rely on technology, making it the heart of most businesses. The rapid development of digital technologies has significantly changed security perspectives and increased the risk of cyber threats. The nature of cyber threats and attacks has changed, and cyber-attacks are now more frequent, complex, and target-orientated, while many businesses still lack the necessary knowledge to defend against them. Researchers are examining attacker...[more]
Morgan Causer	Designing an Interactive Application to Support People Along their Muscle Rehabilitation Journey	Eirini S Anthi	Carolina Fuentes Toro	2223-CM3203
Context: Recently, I was involved in a severe car accident requiring muscle rehabilitation on my back. Further, I found little resources which I could use to help support and educate me during this time. This stretches to my partner, who is currently on a muscle rehabilitation journey (after a fitness injury) for the last several months and has also found it overwhelming, confusing and tedious. Problem: Fitness applications have saturated the app store since coronavirus increased the need for...[more]
Muhammad Huzaifa Jamil	SOC Defense against Dark Art	Yulia Cherdantseva	Amir Javed	2223-CMT400
This research aims to build guidelines for the Security Operation Center analyst in order to respond against different types of attacks. The security operation center is the heart of any organization as it is the first line of defense. SOC consist of mainly four roles Security analyst, security engineer, security manager, and chief information security officer. The security analyst role is further divided into three levels, Our research focuses on all three levels from the detection of the attac...[more]
Nafis Ahmed	Robot Assistant to Prevent and Manage Falls (Robotics + Applied Computer Vision + AI + IoT)	Charith Perera	Alun D Preece	2223-CM3203
Falling is a major risk to the health of older people with approximately one third of community dwelling elderly falling at least once a year (NHS, 2021) and numbers are expected to rise in the UK due to an aging population. Falls are the second leading cause of unintentional deaths worldwide with an estimated 684,000 deaths each year globally (WHO, 2021). Falls occur more often at night and a proposed solution to reduce this was to use a guiding night light which would aid mobility from the bed...[more]
Nicolo Licata	Machine learning for cryptocurrency trading	Yuhua Li	Yazmin Ibanez Garcia	2223-CM3203
Algorithmic automatic trading has always been a hot topic, and it has been studied for a long time. With the development of science and technology, new research results in algorithmic automatic trading emerge. Supervised deep learning and deep reinforcement learning are typical methods. Many researchers proposed many similar or different algorithmic automatic trading with supervised deep learning or deep reinforcement learning. In 2020, Lei, K., Zhang, B., Li, Y., Yang, M. and Shen, Y. proposed ...[more]
Noah Griffiths	Designing an Interactive Application for Adults to Promote Healthy Eating Practices at Home	Nervo Verdezoto Dias	Yulia Cherdantseva	2223-CM3203
There are many reasons for poor healthy eating practices like behaviour and culture, and technology offers a platform that can be used by adults to promote better eating practices (Gov.uk, 2017). Therefore, the aim of the project is to design an interactive application for adults to promote healthy eating practices at home. Firstly, secondary research is conducted to learn how current mobile technologies are used to promote healthy eating practices and how successfully they do so. Secondly, Desi...[more]
Oliver Carter	Data Analytics for Predicting F1 Race Outcomes	Oktay Karakus	Carolina Fuentes Toro	2022-CMT400
Implement a machine learning approach in Python to determine to what extent the results of the 2021 Formula One season can be accurately predicted, based on factors that decided outcomes of the previous year. In order to achieve this, various features of the sport that will be explored to determine how statistically significant they are. Such features will include, but are by no means limited to, starting grid position from qualifying, car constructor, driver and weather conditions. Various Py...[more]
Osama Aloraini	Utilising Technical Analysis, Commodities Data, and Market Indices to Predict Stock Price Movements with Deep Learning	Xianfang Sun	Yipeng Qin	2223-CMT400
The objective of this project is to develop stock price prediction algorithms using two deep learning models: LSTM and CNN. The data features utilised include stock technical analysis, commodities data like energy prices and gold, and key market indices from the U.S. stock market. Additionally, technical indicators are employed to obtain trading strategies, represented as vectors for data features. Moreover, based on features categories, nine experiments were conducted for each model to assess t...[more]
Oscar Russell	Designing an Interactive Application to Promote Healthy Eating among Older Adults at Home	Nervo Verdezoto Dias	Matthew J W Morgan	2223-CM3203
Prerequisites: -Basic knowledge of Human-Computer Interaction and user-centered and/or participatory design methods -Experience developing mobile and/or web applications in several platforms (e.g., Android, HTML5 & Javascript, etc.) -Desirable experience with Spanish (not mandatory) Digital health technologies are becoming increasingly prevalent to support and enhance the everyday experiences of older adults. However, their potential to promote healthy ageing practices has been limited. T...[more]
Otto Hooper	The impact of machine learning in the retail financial market	Oktay Karakus	Bailin Deng	2223-CM3203
The purpose of this project is to present academic findings based on the possibilities of analysing the financial market, utilising machine learning (ML) tools in favour of the retail investor. Retail investors are defined as non-professional investors who trade financial securities in their spare time as a hobby. The information used to make a prediction in the financial markets in this day and age is not scarce, instead, information is widely accessible and free through services such as Yahoo ...[more]
Peter Snook	Technocamps Abacws	Catherine Teehan	Andrew Hood	2223-CMT400
Using Minecraft for Education, create a world that will allow users to explore core computational thinking theories and methods through play. The world should include a range of interactive activities and supporting educational materials that will engage and encourage GCSE students and above to explore computer science as an option at A level or university. The world will need to align with the new 2022 Welsh curriculum for science which includes all elements of CS. Explore and use the complex e...[more]
Puzhuo Liu	Learning to Detect and Recognise Road Signs	Yukun Lai	Frank C Langbein	2223-CMT400
To improve driving experience, computer vision techniques are widely used for providing assistance. The project aims to develop a prototype that implements some machine learning technique to detect and recognise road signs from real world camera footage. Such technology is also an essential component for autonomous driving.[more]
Qiongyu Lou	Text Visualisation with Word Clouds	Yukun Lai	Yuhua Li	2022Place-CMT400
We are now in the era with overwhelming amount of information. To explore information more effectively, visualisation is often used to emphasise different pieces of information based on their importance. The aim of this project is to visualise text as word clouds such that readers may obtain general ideas at a glance. Some typical examples are tag clouds used in Flickr (http://www.flickr.com/photos/tags/) and more general Wordle (http://www.wordle.net/). The project involves implementing an algo...[more]
Raedah Almarzooq	Interactive sign language learning with real-time hand sign recognition	Dr Daniel J. Finnegan	Paul L Rosin	2223-CM3203
British sign language (BSL) is a non-verbal visual language that is considered the primary mode of communication for deaf individuals in the United Kingdom (Jabb,2021). There are several resources for learning sign language from printed materials to online learning websites. Despite the usefulness of these resources as they offer different activities for learners, they all have the same limitation which is the lack of feedback. Receiving feedback and comments is crucial for the advancement and p...[more]
Rahul Kaliambi Raghavan	GPS Mobile App	Crispin Cooper	Padraig Corcoran	2022Place-CMT403
Driving on a narrow lane, you're always in the middle. When the lane joins a wider road, it's important to turn onto the correct side. Paradoxically this is harder to do on an empty road, as there are no visual cues to remind the driver which side they should be on. This project proposes development of a mobile app to remind drivers as they approach situations such as these. This may work best as a plugin to existing open source GPS navigation software e.g. OSMAnd[more]
Rhys Clements	Detecting attacks against IEEE 802.11 Enterprise Networks	Eirini S Anthi	Hiroyuki Kido	2223-CM3203
With the proliferation of smart portable devices like smartphones, tablets, and IoT devices, WiFi (IEEE 802.11) has been established as the dominant technology for connecting digital devices in Wireless Local Area Networks (WLAN). However, the security of WiFi-based networks, as well as the 802.11 protocol itself, have not been studied extensively. This project will use the recently released AWID3 [1] dataset to further understand attacks that may compromise this system and build a detector to p...[more]
Rhys Fourie	Implementation of rhythm-based content generation for games	Frank C Langbein	Amir Javed	2223-CM3203
When playing video games, a player will often encounter a simple problem, and that is that they complete all of the available content within the game. Some games can avoid this issue by focusing on a ‘high score’ system and arcade-styled gameplay loops. However, this is flawed, as in order to retain interest in the game, certain elements are randomized, in order to ensure variety between playthroughs. The issue stems from this: the randomized content does not provide enough reason for repeat...[more]
Richard Jackson	Migration of Nightingale HQ’s website from Hugo to Umbraco CMS	Martin J Chorley	Kirill Sidorov	2022Place-CMT403
The company Nightingale HQ (NHQ) have asked for their company website (https://nightingalehq.ai/) to be migrated from its current tech stack of HUGO and Netlify CMS to a new tech stack, based around Umbraco CMS. The website migration aims to provide NHQ with the current functionality afforded by its current stack, alongside further potential for extensibility and development. As HUGO is a JavaScript-based static site generator, its functionality is largely dedicated towards serving static cont...[more]
Richard Johns	Could AI do a NHS (Audiology) clinicians paperwork to help create a better patient experience.	Paul Goddard	Víctor Gutiérrez Basulto	2223-CMT403
It is often reported by clinicians in the NHS that their least favorite part was the associated paperwork. Paperwork such as appointment summary notes, letters to doctors, onward referrals etc. are incredibly important. NHS Staff are trained to reject the medical model and embrace the biopsychosocial model of care. However with the growing levels of demand faced by clinicians, they are not given enough time to complete all the medical tasks and write notes and also spend time delving into a pati...[more]
Ricky Chu	Implementing a System for 'Tournament' Peer Grading	Richard Booth	Shancang Li	2223-CM3203
Peer grading is an increasingly popular way to grade students' work, especially in MOOCs (Massive Open Online Courses). The idea is that each student grades the work of a subset of the other students, and from this information a final overall ranking of the students is calculated. Intuitively we want the opinion of students that are ranked more highly to carry more weight than those students ranked lower down, since they, presumably, are better qualified to recognise good quality work. What mak...[more]
Roshan Roy	Implementing and Evaluating Deep Neural Networks for CT Segmentation	Yukun Lai	Aric Fowler	2223-CM3203
We have seen significant progress in medical imaging, which provides doctors/radiologists with a powerful tool for early diagnosis of diseases. However, it requires substantial expertise and is also time consuming to analyse CT images to identify and segment tumours. This project involves implementing some deep neural network based architectures and carry out experiments to test their strengths and weaknesses.[more]
Sai Bhargavi Garikapati	Developing a Medication Reminder Application: Integrating Alexa and ChatGPT	Usashi Chatterjee	Xianfang Sun	2022Place-CMT403
This project highlights the development of a medication reminder app utilising React Native, Django, integrating Alexa and ChatGPT. The app aims to improve medication adherence by providing users with automated reminders and personalised assistance. The development process involves leveraging React Native to create a cross-platform mobile application, ensuring compatibility across various devices and operating systems. Django is used for backend development, enabling efficient data manageme...[more]
Saleh Mohamed	Rules Formation for Cyber Resiliency Metrics	Neetesh Saxena	Tingting Li	2223-CMT400
This project aims to do rules formation and test cases for implementing cyber resiliency metrics for industrial control systems supporting cyber resiliency assessment. This work will improve monitoring capabilities and situational awareness.[more]
Sam Thornton	Recovering keystrokes from audio recordings of keyboard emanations using time difference of arrival data and an English language model	Kirill Sidorov	Alun D Preece	2223-CM3203
Audio emanations from keyboards have been found to contain information relating to the typed content. Given several audio recordings of typing, is it possible to recover the text? Using the difference between wavefronts arriving at microphones to build a geometric model, and a simple model of the English language, how much text can be recovered? This method is based on prior research by Asonov and Agrawal, Zhu et al., and Zhuang et al..[more]
Samuel Hall	TrainingTribe: Enhancing Physical Activity and Well-being through a Clan-Based Fitness App	Dr Soumya Barathi	Kirill Sidorov	2223-CMT403
This study examines the influence of the TrainingTribe application, featuring the “Tribe War” functionality inspired by Clash of Clans, on intrinsic motivation and physical activity levels. Following the design and integration of this feature, along with other elements to complete the application, an experimental study was conducted to evaluate the impact of group competition and social comparison by using the application in the context of fitness. A within-subject study involving 8 par...[more]
Samuel Harries	xSE: Measuring the shooting efficiency in Basketball - A Case Study For NBA	Oktay Karakus	Xianfang Sun	2223-CM3203
Technological advancements and breakthroughs in machine learning have led to a recent revolution in the way that certain sports are being analysed, coached, and enjoyed by fans. The most famous of these statistics is the expected goals (xG) metric used in football. This is a metric that gives the probability that a specific shot in football will result in a goal. This metric is expressed as a number between 0 and 1, with 0 being no chance of a goal and 1 being a certain goal. This statistic has ...[more]
Sarah Parker	Does my burger contain horse meat ?	Liam Turner	Kirill Sidorov	2223-CMT403
Meat adulteration, meat fraud, meat substitution has always been a problem, but it has been highlighted in recent years by the ‘Horse meat drama’. As such, part of the testing requirements for food samples is the determination of each meat species contribution within the total meat content (e.g. is my lamb curry actually ‘lamb’, does my burger contain ‘horse’ meat etc). Two of the main approaches to determine meat species are DNA sequencing and quantitative polymerase chain reaction ...[more]
Sauronil Das	Banking Software for Micro Finance Bank Utilising Hand Held Collection Device	Sandy Gould	Alexia Zoumpoulaki	2223-CMT404
Microfinance Client-Server Web Application developed using Django and Deployed in Railway.app. This standalone software application manages user accounts KYC and enables users to make transactions. The admin panel has complete CRUD capabilities, which serves as the focal point for the management system, taking away old pen and paper methods to maintain transactions and calculations.[more]
Sean Berry	Image to Speech System	Bailin Deng	Andrew Jones	2223-CM3203
In this project, you will build a prototype system that converts the text inside an image to a speech. Such a system can be useful for people with visual impairment. The hardware consists of a computer attached to a camera. The system will use computer vision libraries to extract text that is inside an image captured by the camera, and then use text-to-speech APIs to convert the text to audio. The system is expected to be deployed on a PC with a webcam.[more]
Selim Celik	P3: Object detection within buildings (IoT + Applied Computer Vision + AI)	Charith Perera	Frank C Langbein	2223-CM3203
Service robots currently face a significant challenge in accurately and efficiently perceiving their surroundings which is critical for autonomous tasks such as navigation, mapping, and object manipulation. The challenge arises as the robot has to perceive the world through various sensors, including cameras, lidars, and radars. The robot has to integrate all these different sources of information to create a 3D map of the environment, identify objects, and make decisions. This project focuses o...[more]
Sethu Pastula	Virtual Try-On with Deep Generative Networks	Yukun Lai	Frank C Langbein	2223-CM3203
Virtual try-on refers to techniques that virtually put chosen clothing on a given person and generate realistic images. It has significant real-world applications to support online shopping for clothing. In recent years, we have seen significant improvements in technical advances for virtual try-on, largely benefiting from deep neural networks using e.g. generative adversarial networks (GANs). The project aims to build a prototype system to achieve virtual try-on.[more]
Shahmeer Khan	Cloud-Native CI/CD (Continuous Integration/Continuous Delivery) Pipeline for Containerized Applications and Microservices	Andrew Jones	Bailin Deng	2223-CMT400
The goal of this project is to design and implement a continuous integration and continuous deployment (CI/CD) pipeline tailored for cloud-native applications (as state-of-the-art large-scale applications are now more on cloud native applications due to portability and flexability), utilizing containerization and orchestration technologies to streamline the development, testing, and deployment process. This will be done using the latest frameworks and multiple layers of deployment will be introd...[more]
Shiwali Santosh Charjan	Digital Transformation Project with SCNDI Ltd	Catherine Teehan	Matthew J W Morgan	2223Place-CMT404
SCNDI ltd is a tech enabled fitted space making furniture company. We are improving our customer service experience from the start by tracking clients throughout the whole process, including when they are finding us online, contacting us, filling in their requirements, (using our forms), participating in our free design consultations, placing product orders, purchasing products, fabricating of our products, logistics of supply chain, delivery of finished products, installation of products to ...[more]
Sruthi Siddhabhaktula	Findbridge - A language barrier solution for non-native speakers to connect with local native speakers and opt for their services	Amir Javed	Kirill Sidorov	2223-CMT403
Introduction: In today's world, immigration is becoming more and more common, with individuals seeking better opportunities, improved living standards, and higher education. Immigrants face a number of challenges when moving to a new country, one of which is the language barrier. A lack of proficiency in the local language can hinder their ability to perform even the most basic tasks, such as opening a bank account, navigating public transportation, and communicating with doctors or lawyers. The...[more]
Stella Bejide	Interactive 2D Virtual Map	Dr Daniel J. Finnegan	Oktay Karakus	2223-CM3203
I would like to make a 2D virtual map of a hospital for children/young adults who are unable to visit in person. It would be a 2D map plan with interactive tasks along the way for them to complete. There would be a chat function, where they would be able to communicate with each other and share photos.[more]
Tayyeb Rafique	nea, an iOS app to help you regain control of your attention.	Dr Daniel J. Finnegan	Tingting Li	2223-CM3203
Introduction In 2017 Reed Hastings, Chairperson for Netflix, announced that their biggest competitor was sleep.[i] It’s not just Netflix, big tech companies like Facebook, Google, and Amazon, all compete with a user’s free time, family time, and indeed sleep, to maximise revenue. The ‘attention economy’ – profiting by occupying the brain space of the people, has been growing and shows no signs of slowing down.[ii] Advanced AI models are being pointed at users through applications on ...[more]
Theodor Kozlowski	XAI&I: Closing the Accuracy Gap Between Self-Explanatory AI and Black Box Convolutional Neural Networks	Catherine Teehan	Steven Schockaert	2021Place-CMT403
Using the research conducted in the paper “XAI&I: Self-explanatory Ai facilitating mutual understanding between AI and human experts” (Grange et al. 2022), explore (through data analysis) and build upon the algorithms existing methodology to improve accuracy and feature intelligibility. This can be done through incorporating additional layers, additional networks, manipulation of the input data, or other exploratory means.[more]
Thomas Mamoottil John	Educational Web Application Guide for Stocks Trading, Investing and Technical Analysis.	Liam Turner	Jing Wu	2223-CM3203
This proposal outlines the development of a web application, designed primarily to educate beginners about trading and investing in the financial markets. The application will utilise a frontend comprised of ReactJS and HTML, while leveraging Django and Django REST for backend services. Key technologies implemented in this application include the TradingView Charting Library API, which will provide dynamic charting capabilities, and the OpenAI ChatGPT API, deployed for AI-assisted query searches...[more]
Tianyu Hu	English Speaking Practice Application	Matthew Moloughney	Omer F Rana	2223-CMT403
For non-native English speakers, there are a wealth of applications (free-to-use and paid for) which will help the user to read English, Write English and listen to English. It is more uncommon to find applications which allow the user to practice speaking the English language, from phrases they might suggest, or an existing corpus of phrases. Students at Cardiff University who might be non-native English speakers, may feel less confident when speaking or asking questions in English. There m...[more]
Timothy Standen	GPS mobile app to aid driving safety abroad	Crispin Cooper	Padraig Corcoran	2223-CMT403
When driving on a single-track road in a foreign country, it can be hard to remember which side of the road you must drive on once the road ends. This means that when reaching a junction, a driver may drive on the side that they are used to, which could be the opposite side for the country they are currently in. This is especially important on rural isolated roads, which tend to be quieter often single-track roads and have minimal signage. The project aims to create a mobile app that uses GPS t...[more]
Tito Thomas	Final Year – Anime Recommendation System	Fernando Alva Manchego	Yazmin Ibanez Garcia	2223-CM3203
I propose a project to build a web-based application to help people find movies/anime shows that they would enjoy watching. The problem is that many people struggle to find anime or movies that they enjoy watching and don’t know where to start. As a non-anime watcher, it can be difficult to know which shows are right for you and as a result, many people may have a bad first experience. The goal of this project is to build a system to better introduce people into this new style of media that is...[more]
Uday Motiram Sawant	Web Interface and Cloud Deployment for MRSNet	Frank C Langbein	Dr Soumya Barathi	2223-CMT400
MRSNet is a deep learning tool to quantify metabolites in magnetic resonance spectroscopy (see https://qyber.black/mrs/code-mrsnet - more code, by now also ported to python3 with many extensions and clean-ups, will be available if you choose to join the project; we are about to release V2). This is relevant for medical diagnosis as well as understanding of biochemical processes. The aim of this project is to develop a UI to make MRSNet more usable in the relevant application domains and simplify...[more]
Wei Fu	Image super resolution	Xianfang Sun	Yipeng Qin	2223-CM3203
Image super resolution is the techniques of producing high resolution images from low resolution ones. The aim of this project is to analyse and implement some image super resolution algorithms. The student should have a good programming skill and a relatively strong mathematical background.[more]
Wenhao Hong	Image forgery detection with deep learning	Xianfang Sun	Sylwia Polberg	2223-CM3203
Image forgery detection is to detect images that have been modified using some image processing tools. The aim of this project is to analyse and implement image forgery detection algorithms with deep learning. The student should have a good programming skill and a relatively strong mathematical background.[more]
Will Bromham	Estimating the Energy Usage of Code [multiple projects: must be specialised]	Frank C Langbein	Padraig Corcoran	2223-CMT403
Traditional performance evaluators of code are linked to CPU/GPU, memory and IO operations. While there is a link to the power usage of a program, this is quite complex. The aim of this project is to investigate the power usage of code (not hardware, even if the same code may have different power profiles depending on the hardware; we are interested in the energy usage on specific hardware; if you are interested in co-design aspects instead, please discuss feasibility). The first task is to fi...[more]
William Beazley	Using machine learning to determine the best fantasy premier league players	Oktay Karakus	Amir Javed	2223-CM3203
For my proposal, I would like to look deeper into different machine learning models to determine the best possible fantasy premier league (FPL) squad in the upcoming game week based on the players fitness, opponents and previous performances. Fantasy premier league is a web based game which allows users to create a team which they can put out each week. this team will then be scored by how each of their players perform in real life matches. there are many rules that players have to abide by when...[more]
William Payne	Technocamps Islands	Catherine Teehan	Juan Hernandez Vega	2223-CM3203
Using Minecraft for Education, create islands that will allow users to explore core computational thinking theories and methods through play. The islands should include a range of interactive activities and supporting educational materials that will engage and encourage children and young people to explore computer science as an option at GCSE and beyond. The islands will need to align with the new 2022 Welsh curriculum for science. This includes all elements of CS and can be adapted for teach...[more]
Xinyu Liu	Learning and Detection of Road Signs	Shancang Li	Tingting Li	2223-CMT403
Traffic sign recognition plays a crucial role in intelligent transportation systems, and its accuracy and real-time performance directly affect road safety and traffic flow. At present, deep learning based object detection algorithms have made significant progress, with YOLOv5 being widely used in the field of object detection as an efficient real-time object detection framework. However, there are still some challenges in small target detection and target feature extraction of traffic signs. T...[more]
Xuechun Liu	Sentiment analysis of AI-related posts on English and Japanese social media	Alun D Preece	Yuhua Li	2022Place-CMT404
This project conducts a sentiment analysis on AI-related posts on English and Japanese social media. The objectives are to discover trends of AI-related conversations over the years on English and Japanese social media and uncover public attitudes towards AI and AI-related subjects (whether most people think positively, neutrally, or negatively about AI development). This project uses data collected from multiple Mastodon open instances (social media platforms).[more]
Yan Pui Kong	A Proposed Framework for VR Planning and Development in Collaboration with Virtus Tech	Catherine Teehan	Carolina Fuentes Toro	2022Place-CMT404
Virtus Tech is a Virtual Reality software company that creates apps for the web and Quest 2 VR headsets. Their primary way of building apps for the Quest 2 headsets is to use a physics and rendering engine known as Unity that has XR support for VR and AR devices. One of Virtual Reality’s main differences compared to traditional monitor and screen hardware is that it has an extra axis in linear input compared to a mouse and has three dimensional spatial rendering that presents its view relati...[more]
Yanqiao Huang	Image to Speech System	Bailin Deng	Yuhua Li	2223-CM3203
In this project, you will build a prototype system that converts the text inside an image to a speech. Such a system can be useful for people with visual impairment. The hardware consists of a computer attached to a camera. The system will use computer vision libraries to extract text that is inside an image captured by the camera, and then use text-to-speech APIs to convert the text to audio. The system is expected to be deployed on a PC with a webcam, or on a raspberry PI with a camera module.[more]
Yash Ranjan	VR Chess Application	Yukun Lai	Kathryn Jones	2223-CM3203
With the availability of low-cost virtual reality solutions, including mobile phone-based approaches, virtual reality (VR) has become much more accessible to users. However, most chess applications for this platform have problems ranging from UI/UX to the rules of chess not being implemented well. The aim of this project is to create a VR chess application that accurately represents western chess as it is played in real life.[more]
Yifang Li	Collaborative Work Web Site for Research Groups	Alia I Abdelmoty	Juan Hernandez Vega	2223-CM3203
This project will design and build a web site to allow research groups to work collaboratively together. A user-centred approach will be adopted to define the requirements and design and evaluate the project. Beyond Web page design, the site needs to allow researchers to contribute information about their research projects, work together on a shared reference store, keep track and provide alerts to relevant new publicatins/journals/calls. This is also an exercise in HCI and evaluating the usabi...[more]
Yintao Yang	Discovery of the recent music trends based on Sentiment Analysis in Twitter or Spotify	Dr Daniela Tsaneva	Jing Wu	2223Place-CMT403
Music industry is moving fast, and new trends are emerging quickly. This project will work on predicting the new trends in music genres based on Sentiment Analysis in Twitter or Spotify. By analyzing the sentiments of the users automatically we can provide statistical information about the music genre which is favorite to most of them and based on that make a prediction of future music trends. Good Python programming skills is required for this project since a Sentiment Analysis classifier need...[more]
Yitong Yu	Look Ahead: Street View Based Navigation of Roads with Traffic Signs Recognition	Bailin Deng	Juan Hernandez Vega	2223-CMT403
In the UK, the overall pass rate of practical driving tests is only around 50%. One of the top reasons for failing a driving test is "not responding correctly to traffic signs" (see https://www.gov.uk/government/publications/top-10-reasons-for-failing-the-driving-test/top-10-reasons-for-failing-the-driving-test-in-great-britain). In this project, you will develop an application to help learner drivers get familiar with traffic signs on the roads. The app will utilize google street view to allow ...[more]
Yiwei Su	Hearing the future: Predicting the next piece of audio	Dave Marshall	Nervo Verdezoto Dias	2223-CMT400
The basic ideas is based on a sequence of audio can you predict the next few seconds. Deep learning networks (E.g LTSM, Recurrent Neural networks) can be used. Training data is abundant: any audio of a few seconds and be utilised. Take a few second segment and use this build a model that predicts the next segment. A variety of interesting questions need to researched: * What the the best format for the input audio * What type and configuration of network is best * Format of training data: how ...[more]
Yizhou Shen	Learning Effective Security Strategies through Reinforcement Learning	Tingting Li	Ian M Cooper	2223-CM3203
Defending cyber security is a significantly unfair game between defenders and attackers, as defenders need to be cautious all the time to detect and react to every single attack, whilst attackers only need to strike once at any time. Rapid development in Artificial Intelligence (AI) provides the potential for distributed, adaptive defensive measures at machine speed and scale. It is possible now the defender can be trained as an intelligent agent to develop strategies to respond to an attacker a...[more]
Yoshinobu Katayama	Semantic Search using Large Language Models [MULTIPLE PROJECTS]	Alun D Preece	Paul L Rosin	2223Place-CMT404
This project is hosted by the Security Crime and Intelligence Innovation Institute. We are interested in efficient and effective ways of discovering data by semantic search and matching. Semantic matching involves taking into account the meaning of information rather than just it's syntactic structure. Semantically, a "river bank" is very different from a "money bank" even though syntactically they both include the word "bank". Modern natural language processing (NLP) uses large language models ...[more]
Yuzhi Wei	Deep Reinforcement Learning for Game Bots [multiple projects: must be specialised]	Frank C Langbein	Juan Hernandez Vega	2223-CMT403
AI bots in games generally perform quite poorly compared to human performance, without employing tricks to improve their perceived performance. The aim of this project is to investigate reinforcement learning approaches to train human-level game bots, utilising the same or at least similar information than a human gets from the game. You are free to choose the game you wish to try this on and the particular reinforcement learning approach you wish to test (see Doom/VizDoom, OpenAI/DeepMind ex...[more]
Yves Sindayamaze	Managing participants in sailing races	Stuart M Allen	Hiroyuki Kido	2223-CM3203
Many sailing clubs organise weekly races for their members. Although these are typically informal, with sailors simply turning up on the day to take part, it would be useful to keep track of participants for two main reasons. Firstly, from a safety perspective, it would be helpful to know who has taken to the water to race, and more importantly that they have returned to shore. Secondly, it would be helpful to know exactly who is taking part in a race for the purposes of managing results. The ...[more]
Zaria Cameron	Exploring how ubiquitous technologies can support children's coding skills.	Carolina Fuentes Toro	Liam Turner	2223-CM3203
This project focus on disadvantaged children(socio-economic / racial disparities etc..) and how their disadvantages affect their access to technologies.[more]
Zihao Han	VGG(CNN)-Based Neural Network Optimization for Breast Ultrasound Classification: Benign, Malignant, and Normal	Hiroyuki Kido	Kirill Sidorov	2223-CMT404
In recent years, the application of Convolutional Neural Networks (CNNs) has garnered considerable attention in the realm of medical imaging, given their innate prowess in image representation. The goal of this project is to find a specific medical image dataset, implement neural network models, and evaluate their performance. The project will also involve a critical review of existing models to motivate the work and justify the proposed models.This study seeks to explore the application and opt...[more]
Ziqi Li	Machine learning of where people look in images/videos	Hantao Liu	Irena Spasic	2223-CMT403
Do you know where you are looking at while watching online videos? Do you know where the computers think you are looking at? Modelling visual saliency - predicting where human eyes pay attention to in visual content - has been a very active research area over the past few years in both academia and industry. This project aims to analyse visual attention data (both ground truth and predicted) in emerging applications in image and vision computing, such as computer rendered images, high-dynamic-ra...[more]
Ziqi Wu	Optimizing Cost Efficiency in MLOps	James Osborne	Jose Camacho Collados	2223Place-CMT403
The advent of MLOps provides a structure for automating machine learning workflows, leading to increased efficiency and consistency. However, the computational demands of large-scale model training and deployment can result in significant infrastructure costs. On the other hand, existing solutions are designed to the large models for machine learning, and data scientists and engineers are clearly unconscious of the difference of real cost of training and deploying small models compared with the ...[more]
Zizhou Liu	A Visual Tool for Generating and Back-testing Algorithmic Strategies Under the Chinese Stock Market	Dr Daniela Tsaneva	Federico Liberatore	2223-CMT403
The goal of This project is to design a tool that can help traders with zero programming knowledge to generate automated trading strategies. One of the functions of this tool is to assist with visual trading strategy reproduction, by providing a variety of technical indicators to selectively click on the way, and gradually guide individual traders in the tool flow chart box to build and reproduce their own trading logic flow chart. This way the tool will help individual traders to visualize thei...[more]