Research Degrees

These are the opportunities currently available within the Digital Imaging Research Centre (DIRC).

Funded Scholarships

Fees-only Scholarship (Home/EU)
"3D Protein Structure Matching based on advanced descriptors", PhD Supervisor: Dr J.-C. Nebel Since many of the new 3D protein structures delivered by high-throughput processes do not have any known function, there is a need for structure-based prediction of protein function. Moreover, to date no system allows the fully automated classification of 3D protein structures. We propose the usage of advanced protein family descriptors represented by low dimention structure manifolds. Candidate should have very good programming skills. Knowledge in neither bioinformatics nor biology is necessary. More information and how to apply Supervisor's homepage and their other projects
"Genetics-inspired framework for video processing", PhD Supervisor: Dr J.-C. Nebel Although state-of-the-art visual surveillance algorithms have been the product of extensive work by many researchers worldwide for more than a decade, current approaches are still not sufficient to handle the very wide range of data exhibited by CCTV videos. Consequently, a new approach is required to deal with such variability. In this project, we propose a genetics-inspired framework which sees variability in CCTV data as the expected norm rather than an inconvenience to control. By representing a CCTV sequence as an image under evolution, we will able to design a new range of video processing algorithms inspired by genomics approaches. Excellent programming skills and good knowledge of standard video processing algorithms (foreground extraction, optical flow estimation, tracking, etc.) are required. Familiarity with dynamic programming and machine learning techniques is desirable. More information and how to apply Supervisor's homepage and their other projects
"Efficient “lock-and-key” 3D matching techniques for rational drug design", PhD Supervisor: Dr J.-C. Nebel Rational drug design has revolutionised the development of new drugs. In particular, it relies on solving the “lock-and-key” problem where one is interested in finding the correct, if any, relative 3D orientation of the “key” – potential drug - which will open up the “lock” – protein involved in disease. To tackle this problem, also known as docking, a novel approach will be investigated where “locks” will be represented by advanced binding site descriptors. The aim of this project is, first, to produce 3D binding site descriptors, and secondly, to develop advanced 3D matching techniques which would estimate if and how “keys” could open up “locks”. Such technology would lead to rapid screening of drugs which could potentially be used to inhibit proteins involved in diseases. Candidate should have very good programming skills. Knowledge in neither bioinformatics nor biology is necessary. More information and how to apply Supervisor's homepage and their other projects
"Development of protein binding site descriptors for detection of protein-protein interaction sites", PhD Supervisor: Dr J.-C. Nebel Proteins are large molecules composed of chains of amino acids, which are called protein sequences. These chains adopt three-dimensional (3D) shapes, or structures, which define protein function. Since 3D structures of most proteins are unknown, scientists have investigated the prediction of a protein function directly from its sequence. Since many proteins only become functionally active if they interact with other proteins through specific regions, i.e. binding sites, the detection of these sites could be key to protein functional annotation. The aim of this project is, first, to produce computational tools to analyse known 3D structures of interacting proteins, and secondly, to generate binding site sequence descriptors which could be used to infer binding site location from a protein sequence. Such technology would lead to rapid screening of proteins involved in diseases and the rational design of new drugs targeting specific proteins. Candidate should have very good programming skills. Knowledge in neither bioinformatics nor biology is necessary. More information and how to apply Supervisor's homepage and their other projects
"Development of machine learning algorithms for protein-protein interaction prediction", PhD Supervisor: Dr J.-C. Nebel The scientific community is currently engaged in the ambitious task of identifying all proteins present in the human body. It is expected this project will revolutionise drug design for a whole range of diseases including cancer and diabetes. However, this will be achieved only if better computational tools are available to address the problem of automatic prediction of protein-protein interactions which play a fundamental role in many diseases. The aim of this project is to develop machine learning techniques to predict protein-protein interaction pairs and interaction sites. Learning from known protein-protein interaction networks, specific features, such as gene co-location and co-evolution patterns, will be identified to build classifiers able to differentiate between interacting and non-interacting protein pairs. Refined by the analysis of known protein complex 3D structures, this methodology will also be used to identify specific interaction sites on protein surfaces. Candidate should have very good programming skills. Knowledge in neither bioinformatics nor biology is necessary. More information and how to apply Supervisor's homepage and their other projects

Research Project Ideas

Project List
"Action Analysis for Computer Games", PhD Supervisor: Dr V. Argyriou Computer games have attracted a wide market over the past few years and their appeal may be further enhanced through vision based inputs. The high speed and low cost requirements make system design challenging. In the upcoming generation of computer games interaction will be mainly using human gestu... More information and how to apply Supervisor's homepage and their other projects
"Enhancing game-play through real-time computer vision", PhD Supervisor: Dr D. Greenhill The announcement of the Xbox 360’s controller-less peripheral Natal and Sony’s Motion Controller has generated huge interest as vision based gaming becomes an increasingly important research area. Computer vision control allows the movement or gesture of the player to affect the gameplay by crea... More information and how to apply Supervisor's homepage and their other projects
"Development of 3D image vision approaches for unsupervised high content screening of bioimaging datasets", PhD Supervisor: Dr A. Hoppe The combination of light microscopy and image processing has led to the development of more objective, quantitative analysis techniques in experimental biology. The analysis of biomedical image data is challenging but has provided an insight into complex biological processes. The current problems w... More information and how to apply Supervisor's homepage and their other projects
"Finding Occlusion Layers in Video Streams using Pixel Motion", PhD Supervisor: Prof G.A. Jones Computing the optical flow (pixel motion) requires spatial and temporal pixel regions of support. Such regions may contain motion discontinuities which can be identified using robust statistical approaches. These statistical techniques require initialisation by motion clues from other techniques such as clustering or feature matches (or estimates from previous frames). More information and how to apply Supervisor's homepage and their other projects
"Converting Films to 3D Movies (3DMovie)", PhD Supervisor: Prof G.A. Jones The ground-breaking film Avatar has encouraged movie studios to release 3D versions of classic sci-fi and fantasy films such as Star Wars, The Matrix and The Lord of the Rings. The process of adding depth to non-CG elements in a movie frame, however, is incredibly time-consuming and hence expensive.... More information and how to apply Supervisor's homepage and their other projects
"Cooperating communities of PTZ sensors (CoPs)", PhD Supervisor: Prof G.A. Jones The CoPs project aims to develop the collaborative architecture, visual-ocular reflexes and cross-sensor competencies to allow a community of autonomous pan-tilt and zoom (PTZ) sensors to monitor wide-area scenes. Specific wide-area tasks could include the collection of mug-shots of individuals ente... More information and how to apply Supervisor's homepage and their other projects
"Tracking 2D Articulate Motion", PhD Supervisor: Dr D. Makris Interpreting human motion can benefit a wide variety of industrial sectors, such as sports analysis for virtual replaying and analysis of athletes’ motion patterns, visual surveillance for gait analysis, behaviour understanding and event detection, Human Computer Interaction for body motion based ... More information and how to apply Supervisor's homepage and their other projects
"Active Visual Surveillance", PhD Supervisor: Dr S.A. Velastin This project refers to the use of pan-tilt-zoom (PTZ) cameras (or very high resolution cameras) for actively detecting and focusing situations of interest in the context of intelligent monitoring of spaces not only for security but for more "benign" applications such as assisted living (elderly, disabled). More information and how to apply Supervisor's homepage and their other projects
"Action recognition in cluttered environments", PhD Supervisor: Dr S.A. Velastin This project focuses on being able to understand body language, its dynamics and the interactions between people with a particular focus on understanding actions in environments which are cluttered either because of the number of people or restrictions in the environment, for example when using on-board cameras in public transport. More information and how to apply Supervisor's homepage and their other projects

Centre Director

Velastin S.A. Dr