About Research Publications Reports CV Contact

Dr. Paul A. Bromiley

About me

I am currently employed in the Centre for Imaging Sciences at the University of Manchester. Prior to this I studied for my Ph.D. at the Mullard Space Science Laboratory, a department of University College London, where I worked on the development of adiabatic demagnetisation refrigerators for use on astronomical satellites. I also hold an M.A. in Natural Sciences (Physics and Theoretical Physics) from University of Cambridge and an M.Sc. in Astrophysics from Queen Mary University of London.

Publications

My publications page provides access to my published papers. An alternative list is maintained by the University of Manchester. You can also visit my Google Scholar page.

Reports

A large collection of my unpublished technical reports can be obtained from the TINA Memo System. These cover a variety of subjects in statistics, medical image analysis, machine vision, and the open-source TINA machine vision software itself. In particular, Products and Convolutions of Gaussian Distributions has proved extremely popular, with 40 Google Scholar citations to date.

Some of my past lectures are available here.

Short CV

Employment

2000-present: Research Associate/Fellow, Centre for Imaging Sciences, Institute of Population Health, University of Manchester.

My research career has focused on the development of novel machine vision algorithms based on quantitative statistical techniques, primarily for medical applications. This involves programming in C/C++ on both Windows and Linux platforms, integrating algorithms into both external and in-house software libraries (VXL, CLUE and TINA). I have also been the web-master of the TINA website since 2003.

I have published over 40 refereed journal and conference papers in the areas of machine vision and medical image analysis, and act as a reviewer for a number of national and international journals and conferences.

Selected Projects:

2016-present: STOpFrac: A Software Tool for Opportunistic Diagnosis of Vertebral Fractures.
2013-2016: An Automated Tool to Identify Vertebral Fractures in Various Imaging Modalities.

These projects, supported by the NIHR through the i4i scheme (2016-2019), and the Wellcome Trust and the Department of Health (2013-2016), are collaborations with the Central Manchester NHS Foundation Trust and Optasia Medical. The aim is to develop and commercialise both fully and semi-automatic software packages that identify and measure osteoporotic vertebral fractures. I have:

  • Developed appearance models of the spine, achieving state-of-the-art accuracy, and techniques to fit them to clinical images without a-priori knowledge of which vertebrae are present;
  • Delivered libraries and an API to Optasia Medical, supporting the development of radiological software that allows semi-automated annotation of vertebrae on medical images, and measurement of fractures;
  • Managed a team of five radiographers and a radiologist who perform data acquisition and image annotation.
2012-2013: Monitoring the Faces of Drivers.

The aim of this long-term collaboration with Toyota Motor Europe is to develop face detection and tracking software, supporting driver monitoring for safety purposes.

2012-2013: Tools for Detecting Patterns of Facial Behaviour.

The aim of this EPSRC Knowledge Transfer Agreement with Bunnyfoot Ltd. was to develop an automated tool to identify emotions in videos of faces, recorded whilst the subjects were watching advertisements in various media, allowing monitoring of viewer engagement. During this project I:

  • Developed RFRV-CLMs for facial landmark annotation, supporting classification of emotional states.
  • Developed an application to support manual and automatic annotation of emotional states on videos of faces.
2009-2012: Automatic Identification of 3D Landmarks in Micro-CT Mouse Skull Data.

The aim of this collaborative project with the Max Planck Institute for Evolutionary Biology was to develop software allowing the manual and automatic annotation of morphometric landmarks in 3D medical image volumes, to support genetic studies of rodents. During this project I:

  • Developed and deployed a new version of the TINA software (ver. 6) incorporating a Gtk+ interface;
  • Added 3D volume rendering capabilities to the TINA software, using the OpenGL-style Volpack library;
  • Delivered a ~50,000 lines-of-code software package for manual and fully automatic landmark annotation.
2005-2009: University of Manchester "Stepping Stones" Fellowship.

Application for this scheme was competitive, with an acceptance ratio of 1/5, and involved the generation of a research proposal by the candidates. Through this award I brought funding worth £22,500, plus my own salary, into my research group. During the project, I developed techniques to incorporate spatial information into medical image segmentation and registration algorithms without the use of spatial prior probabilities.

2001-2005: Medical Images and Signals Interdisciplinary Research Collaboration (MIAS IRC).

The MIAS IRC was a collaborative project between Manchester University, Oxford University, King's College London, University College London, and Imperial College London, and was the U.K.'s flagship medical imaging research project. During this project I developed a technique to quantify errors in mutual information image registration, through a novel theoretical interpretation of the algorithm. I also developed a diagnostic decision support algorithm for a range of dementing diseases, based on measurement of CSF volume from MR images, using a novel genetic algorithm optimiser.

2000-2001: SIMBAD (Smart Inactivity Monitor using Array Based Detectors).

The aim of this collaborative project with Infrared Integrated Systems Ltd. (IRISYS), the University of Liverpool and BT plc was to develop an infra-red camera based monitoring system for sheltered accommodation units to detect falls. During this project I developed both non-parametric image subtraction, an image differencing algorithm, and a colour image segmentation technique based on density estimation in colour spaces.

Education

1995-2000: Mullard Space Science Laboratory, Department of Space and Climate Physics, University College London. Ph.D. in Physics "Development of an Adiabatic Demagnetisation Refrigerator for use in Space".

My Ph.D. involved the design, construction, and testing of two novel adiabatic demagnetisation refrigerators in collaboration with Oxford Instruments and Cambridge University.

1994-1995: Department of Mathematics, Queen Mary University of London. M.Sc. in Astrophysics.

1991-1994: Corpus Christi College, Cambridge University.
M.A. in Natural Sciences (Physics and Theoretical Physics).

1984-1991: Valentines High School, Redbridge, London.
A levels: Mathematics (A), Physics (A), Chemistry (A), Biology (A).
ABRSM Grade 8 Classical Guitar and Grade 7 Music Theory.
GCSEs: 9 A grade GCSEs, including Mathematics and English.

Skills

  • 17 years experience of C/C++ programming under Linux and Windows; familiarity with GCC and associated software (CMake, GNU Autotools, DDD/GDB, Valgrind etc.), MS Visual Studio, Qt and Gtk+ interface design, parallel programming with OpenMP, and OpenGL-style rendering libraries.
  • 5 years experience of Python programming, and data analysis and visualisation using SageMath.
  • 14 years experience of web site development and maintenance, including familiarity with standard Linux software (Apache etc.) and software for open-source project management (Git, CVS, LXR, Doxygen etc.).
  • 23 years experience of data analysis and visualisation using Mathematica.
  • 17 years experience of Linux installation and system management, and 12 years of experience of live-Linux development (see the TINA Knoppix page).
  • 17 years experience of PC system integration, including construction of complete systems.

Teaching

I have lectured on medical image registration and colour image analysis on the Advanced Machine Vision Module of the University of Manchester M.Sc. in Advanced Computer Science, and also ran laboratory demonstrations for the course. I have also been partly responsible for the production, and solely responsible for the deployment and maintenance, of the related web-based teaching materials. Finally, I have lectured on the statistical foundations of medical image analysis for the MIAS IRC (Medical Images and Signals Interdisciplinary Research Collaboration) Summer School.

I have been official co-supervisor for two Ph.D. students, and currently unofficially co-supervise one student at the Univeristy of Manchester, and assist with the supervision of another at the University of Sheffield.

I also act as a Tutor for Personal and Professional Development (PPD) for ten Year 1/2 MBChB students in the School of Medicine at the University of Manchester. I assist the students in the development of their PPD portfolios, covering issues such as the development of independent learing techniques, maintaining records of their clinical skills, the use of refelctive writing, and issues related to probity and good medical practise.

Funding

2005-2009: University of Manchester "Stepping Stones" Fellowship.

Application for this scheme was competitive, with an acceptance ratio of 1/5, and involved the generation of a research proposal by the candidates. Through this award I brought funding worth £22,500, plus my own salary, into my research group.

2016-2019: STOpFrac: A Software Tool for Opportunistic Diagnosis of Vertebral Fractures.

I was a major contributor to, and named researcher on, this successful application to the NIHR i4i scheme, bringing funding worth £866K into my research group.

Professional Memberships

1994-present: Member of the Institute of Physics

2000-present: Member of the British Machine Vision Association

2004-present: Member of the Medical Image Computing and Computer-aided Intervention (MICCAI) Society.

Contact

Dr. Paul A. Bromiley
Centre for Imaging Sciences
University of Manchester
Stopford Building, Oxford Road
Manchester M13 9PT, U.K.
Tel: +44 (0)161 275 5175
E-mail: paul.bromiley@manchester.ac.uk