Skip to content

Example research and industry projects

Satellites: Processing satellite images

Our research aims to create a computer application environment for predicting the solar activities (i.e., flares and coronal mass ejections (CMEs)) that could have severe effects on life on Earth.

Solar activity can cause severe problems for the space industry; earth based electromagnetic communications and power systems, radio transmission and so on. An image-processing and machine learning environment will be developed to provide short-term prediction for the occurrence of these events, based on the data that is available in the public domain for other solar features.

We have recently developed a system called "Automated Solar Activity Prediction". This system uses MDI Continuum and Intensity gram images to predict solar flares.

Healthcare: From cemetery to clinic

Digitised pathological data from archaeological leprous skeletons

The project was based at the University of Bradford, and is the result of a fruitful collaboration between Archaeological Sciences and the Centre for Visual Computing.  It was led by Dr Andrew Wilson, and including Dr Jo Buckberry, Professor Hassan Ugail, Dr Christopher Gaffney, Dr Christopher Watkins, Mr Andy Holland, Mr Tom Sparrow, and Dr Rebecca Swift.

Leprosy is a debilitating disease with a strong social stigma. Once common throughout the old world, it is still encountered in the developing world. The project aimed to digitise data pertaining to leprous medieval skeletons and clinical x-rays of modern sufferers in order to allow medical historians, palaeopathologists, clinicians and the interested public to observe and better understand the skeletal lesions of the disease, how they manifest across the skeleton and how they arise. This links to another aim of the project which was to inspire an emotional response from the wider public. This was achieved by offering people the change to come face-to-face with the realities of the disease and it provided the opportunity to gain insight into how people in the past may have responded to the social stigma of the disease.

3D laser scans of affected bones from the medieval leprosarium of St. James and Mary Magdalene, Chichester, alongside digitised x-rays of these individuals and those of clinical cases were disseminated via an interactive GIS cemetery plan and associated database, with pathological descriptions. The project required rapid digitisation to ensure the sustainability of this unique, but extremely fragile collection.

Healthcare: Predicting how tablets crush

One way to improve the design of pharmaceutical tablets is to understand how they'll break down under physical pressure. In this study, working with another University department, we created a computer simulation to model how different sizes and shapes of tablet behave when crushed.

The aim was to develop a computer based simulation platform to study the crushing behaviour of Pharmaceutical Tablets working with the Institute of Pharmaceutical Innovation, University of Bradford.

Using mathematical modelling we developed a computer platform that can predict the crushing behaviour of a typical pharmaceutical tablet. The computer model has been tested and validated using real data supplied by the Institute of Pharmaceutical Innovation.

The platform enables to predict the crushing behaviour of tablets and would enable better design of tablets.

 

Healthcare: Automated analysis of digital mammograms

Breast cancer screening is currently carried out by radiologists, who have to individually examine each result. What if an automated computer analysis could accurately detect abnormalities instead, cutting workload and speeding up diagnosis?

A Computer Aided Diagnosis (CAD) is designed to reduce the workload of radiologists by offering automatic detection, verification and/or classification of abnormalities in digital mammogram images such as microcalcifications, circumscribed masses and speculated lesions.

The development of a computerised CAD system with advanced features to accurately detect the abnormal structure in mammograms and to provide accurate measurements of the size and the location of tumours.

Healthcare: Analysing 3D images of skin and wounds

This provides advantages to both health care workers and patients if damaged skin can be measured and evaluated automatically, non-invasively, using a 3D visual recognition process.  The aim was to develop a computer-based platform for accurately measuring the properties of skin and wounds for Bradford Royal Infirmary (Proof of Concept).

We developed a non invasive computer-based platform that can analyse 3D images of skin and wounds, in order to characterise both the skin and the wounds. In particular, based on the information provided in 3D images we can automatically calculate the surface area and volume of the wound.

The system can accurately undertake essential measurements of skin and wounds to enable health care workers to evaluate the available therapies.

Healthcare: Intelligent Analysis of EEG signals

The Advanced Digital Institute (ADI) developed a new system for measuring biosignals from the heart ( ECG, or electro cardiogram), muscles (EMG, or electro-myogram), the brain ( EEG, or electro-encephalogram) and the gavlanic skin response (GSR).

The main goal of this project was to understand and extract patterns of the sensors, the EEG in particular, and understand actions that affects this signal in order to quantify awareness.

Thus far this collaboration has developed novel processing and machine learning approaches to analysing and classifying the biosignals generated by ADI's biosensors.

Healthcare and security: Predicting human facial ageing

There are plenty of applications for the science of accurately forecasting how human faces will look after years of ageing. It's not just the passage of time that can be simulated: health care professionals can also use our platform to demonstrate how the changes are excpected to differ with a variety of lifestyle factors.

The aim was to develop a computer based simulation platform for simulation of human facial ageing based on 3D scan data of faces. It was funded by Engineering and Physical Sciences Research Council (EPSRC).

We developed a computer based platform for prediction of human facial ageing. Based on 3D scan data of a given face human face our platform can predict the age of the person and can predict the facial changes as the person ages. The system can also predict the facial changes due to other external factors such as smoking and use of drugs.

Security: Building a passive lie detector for UK border control

Being able to use a passive lie detector during interviews would greatly assist UK security agencies. This project set out to analyse visual and thermal images to find out whether such a system would provide accurate results.

The aim of this project to develop a lie detector during an interview scenario based on the analysis of human facial expressions, eye movement and micro facial expressions based on video and thermal images for border control applications. Both visual cues (based on facial images through video) thermal cues (based on images from thermal cameras) combined with psychological analysis provide clues where a person may not be telling the truth.

Partners included EPSRC, UK Border Agency, QinetiQ (project in partnership with Department of Computer Science, Aberystwyth University).

Security: 3D face recognition for security systems

We worked with Yorkshire Concept to develop an efficient system for recognition of human faces in 3D. We developed a system in which we can efficiently compress captured scan data of human faces. We then developed a number metrics based on the geometry of human face (e.g. nose tip, symmetry profile and eye profile).

We have developed a prototype system which demonstrates that 3D facial data can overcome many inherent problems in image-based (2D) face recognition. This can be accurate up to the level of differentiating between identical twins.

 

Visual media: Compressing 3D data For online computer games

A current challenge in online computer gaming is how to transmit data-hungry 3D images using existing broadband networks. Is it possible to maintain realistic levels of detail while reducing the quanity of data required?

The new techniques we developed during this project have resulted in two European patents and a spin-off company with venture funding.

The aim was to demonstrate that geometric data in computer games can be efficiently compressed and rendered using mathematical functions based on solutions to Partial Differential Equations (PDEs) working with Enterprise Ventures Ltd.

Based on patented work, Enterprise Ventures invited us to demonstrate that existing geometric game data can be efficiently compressed. This was part of a feasibility study and due diligence to identify if the patented technology has potential for a future market product.

The problem is that current approaches for representing complex 3D models are not well suited for use in web browser applications. They either require significant bandwidth to download a detailed 3D object, or sacrifice detail to enhance response speed for the end user. Although internet bandwidth is slowly increasing, a software approach could make data distribution easier for a wide range of industries.

The techniques developed through this project allow the actual game content to be compressed by 90% compared to existing techniques.

The project resulted in a spin-out company, Tangentix Ltd.

Visual Media: A virtual reality design tool

This EU-funded project shows how multimedia and virtual reality telematics can be combined into an effective tool for the design industry.  The aim was to work for a number of design services in Europe on the viability of multimedia and virtual reality telematics as an economically justified and professionally effective tool for the design industry. 

The project attracted €4.8 million from the European Union and involved 15 partners designing and implementing the infrastructure to support online design services. 

The result was a Design Information Centre and a Design Services Centre were implemented for target users as a prelude to commercial exploitation. The system handled conventional multimedia data assets (including video), 3D CAD data and simulation.

Visual media: Visually accurate simulations using free rendering software

 

This research study demonstrated a new way to create physically accurate visual stimuli for the study of human visual perception using free software developed for lighting design.  The aim is to establish if RADIANCE, freely licensed ray-tracing software, could be used to produce stimuli for the study of human perception. The project was funded EPSRC (Engineering and Physical Sciences Research Council).

RADIANCE is a physically based, freely available, and commonly used rendering software. We modified the program to allow for spectral rendering and validated the calculation accuracy by comparing simulation results with measurements from real scenes with particular emphasis on colour reproduction.

At the time of publication there was no better simulation alternative for achieving physical realism (i.e., the simulation result providing the same visual stimulation as a real scene) than by combining our spectral rendering method with RADIANCE.

Visual media: Virtual TV broadcasting and presenters

We are using visual media techniques to create virtual environments to be used in TV mediums:

Broadcasting from a virtual studio – using €3.6 million from the European Union and working with a number of European broadcasters and content creators we brought them together with users into an interactive virtual studio environment supported by broadband networks. The project was codenamed VISTA (Virtual Interactive Studio Television Applications using Networked Graphical Supercomputers).

The application was an interactive 10 episode television drama series where each episode combined interactive virtual environment techniques with the television medium.

Interactive presenters – Worked with BBC, Sony and Advance Multimedia on a £752,000 project funded by DTI and EPSRC to develop a talking virtual presenter that addressed the user through a text-to-speech system making appropriate comments and suggestions to assist the user in the selection process

Virtual rehearsal environment - Working with the National Film and Television School with funding of £440,000 to create an image and video creation and manipulation for artificial environments for rehearsal and production of a scenario in a distributed virtual environment.

Visual media: Creating a virtual reality amusement park

 

The European Commission-funded €1 million for the VPARK project aimed to integrate several applications based on Distributed Virtual Environments.

Partners included Integrated Information Systems s.a.(Greece), NMPFT (UK), University of Geneva (CH), and Swiss Federal Institute of Technology (CH).

Attraction builder software, and four key attractions: Virtual Dancing, Virtual Rock Concert, Virtual Theatre, Virtual Olympic Games. 

The theatre application allowed physically distributed actors to link into a common virtual stage and practice their lines and movement before travelling to the physical theatre for the final dress rehearsal.