Benjamin Irving








About me:

Researcher and developer in the intersection between medical imaging, mathematical modelling and machine learning, working at the University of Oxford.

I am currently working on methods for characterising tumours from medical images to improve personalisation of treatment, in collaboration with radiologists and oncologists from leading UK hospitals. Previously, I've worked with exciting startups in the trauma imaging, mamography and computer assisted detection fields. I obtained a PhD in medical imaging and computer assisted detection from UCL in 2013.

Love Python and regular user of c++. Experienced in machine learning and data analysis. Exploring deep learning.


Selected recent projects

Computer assisted detection of subregional variation in tumours
[in progress ...]

Automatic tumour segmentation
Development of methods for automatic and semi-automatic segmentation of colorectal tumours using subregional supervoxel analysis.
Image from 1361-8415/ 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (
see [Publications]

Functional image viewer and analysis for cancer research
Development of a software platform for modelling and analysis of functional images, aimed specifically at DCE-MRI. Used by engineers and clinicians affiliated to the University of Oxford for research purposes.

[Link to viewer documentation]

Automated analysis of airway pathology
see [Publications]


Machine learning, pattern recognition and statistical analysis
5+ years working on machine learning and statistical analysis applications including feature extraction, clustering, supervised and unsupervised learning, deep learning, structured learning, Markov random fields and parts-based graphical models, and assessment of these models using validation and significance testing. Experience with matlab and python ML and statistical toolboxes including scikit-learn, scikit-image, pandas and theano.

Software development and prototyping
Experienced in both prototyping methods and software development in Python, C++ and Matlab. Including Qt GUI development, unit testing and git version control. I also have some knowledge of html and css.

Data visualisation
A key part of my work is presentation and visualisation of data, including 3D mesh models, heat maps and regular plots. I've gained experience working with data visualisation from Matlab, Matplotlib, Mayavi and Pandas, and a working knowledge of X3dom for 3D visualisation on the web.

Medical image analysis and computer vision
Collaborations with leading clinicians to develop algorithms for improved assessment of medical images in order to personalise patient assessment. I've published in leading conferences and journals in the field, such as the journal of Medical Image Analysis and MICCAI. I routinely use pydicom, scikit-image, nibabel and nipy, and have a working knowledge the dicom standard.

Mathematical modelling
Having a background in physics and applied mathematics, I've developed a number of mathematical models with applications in imaging physics and therapy, which have been used in both industry and academia.

Multidisciplinary collaboration
I've worked in both industry and academia in multidisciplinary research and development teams.

Team and project leadership
Experience supervising and guiding projects that interface imaging and the clinic.