Research Computing and Imaging

Today, images are everywhere and are increasingly used in science. There are many examples of the importance of images over the history of science, a classic example being the motion sequence photographs of Eadweard Muybridge taken in 1878. These showed for the first time that for a brief moment when a horse gallops none of its feet are on the ground, thus capturing the truth through photography and paving the way for the development of the motion-picture industry ten years later.

Imaging science is the multidisciplinary pursuit concerned with generating, collecting, analysing, enhancing, and visualizing images. Despite the UK being strong in imaging and having many strategic centres of excellence supporting a variety of imaging methods it currently lacks integration with research computing, which is essential for processing and understanding the detail of these images. These facilities are traditionally based around one imaging method but this fellowship copies the SCI (Scientific Computing and Imaging) Institute in Utah which is an international centre of excellence where software is developed across the range of imaging methods and for the variety of disciplines that use images. This approach allows knowledge to be transferred and software re-purposed across a variety of imaging communities. In the long term, by major goal is to develop a centre of excellence at UoL similar to the SCI Institute.

There are two strands to the fellowship; the first is to develop software while the second is to upgrade computational skills and develop a computational community for imaging.

In this fellowship I will work with two of these national imaging facilities to develop software for a number of strategic case studies. Seven of these will based around new spectral imaging methods along with an initial training case in super-resolution light imaging. The new spectral imaging methods are novel because they include spectroscopy information for each pixel or voxel of the 2D or 3D image. The fundamental physical principles for extracting the exact structural chemical information in spectroscopic X-ray and electron micrograph (em) images are now well established, but there are no software packages that implement the whole of this computational workflow. This fellowship will develop easy-to-use software for 2D, 3D and 4D visualization of X-ray and em images, integrating software tools, re-purposing algorithms and designing new visualization techniques.

This fellowship will also upgrade computing skills in imaging at UoL. I will do this by working with a stakeholder group which includes four senior researchers. The PDRA and I will develop software for them, and those they collaborate with, and I will mentor one of their researchers who support others in their research groups. Knowledge from this mentoring will inform me about what training materials need to be developed on campus. I will also lead three computational networks at the UoL which will be advertised across the N8 (Northern 8 Universities). There will be one meeting a month across these networks which includes an annual one day conference for each of the networks.

While working on these two objectives I will continue to develop professionally and further develop my research into the role of the RSE. With help from the SSI I will disseminate this research and campaign for the RSE role both at UoL and nationwide.

Imaging is increasingly used in science because it is a leading technology platform that enables discoveries in many disciplines. The UK has a number of internationally recognised centres of excellence which are strategically important to the UK economy and society, pushing back the frontiers of our scientific knowledge. This fellowship partners with 2 of these facilities: DLS which is a large synchrotron facility and SuperSTEM which is the EPSRC National Facility for Aberration-Corrected Scanning Transmission EM, which has a range of academic and commercial partners aligned with EPSRC remit, including Computing, Mathematical Sciences, Physical Sciences, Engineering, Digital Technologies, Energy, Healthcare Technologies, and Manufacturing.

During this fellowship a series of software will be released, developed through strategic case studies that support the development of the new spectral imagining methods. While the fundamental physical principles for extracting exact structural chemical information in spectroscopic X-ray and em images are now well established, there are no existing software packages. This is typical of imaging which lacks the integration of imaging with research computing, so contributing to the lag in the adoption of new imaging methods and perpetuating a divide between experimentalists and computationalists, even though it is increasingly important to integrate the approaches.

The series of software I will release are relevant to particular industries in the UK and the design and development of new products as the case studies build on established collaborations with industry, specifically Infineum, AstraZeneca, Syngenta and Saudi-Aramco as well as with academics who also collaborate with industry. These are: 1,) the development of additives to fuels and lubricants; 2,) the formulation and manufacture of prescription pills; 3,) the design of artificial joints; 4,) improved understanding of CFD in the design of a variety of products such as nasal sprays, injections or relevant to the manufacturing of tablets, coatings or printing; 5,) the development of medical devices and tapes/seals; 6,) the development of optical coatings such as anti-reflective, self-cleaning, electrochromic, mirrors, filters and protective, aesthetic or decorative coatings; and 7,) chemical catalysis which is important to the Pharmaceutical, Agrochemical, Fine and Speciality Chemical Industries.

Software will be designed and released through DLS and SuperSTEM and integrated into their roadmaps. DLS has two relevant toolkits, (DAWN) which is designed for data as it comes straight from the beam line, and SAVU which is for the general processing of data. Software will be released into these so as to benefit from the large user communities of both academics and industrial partners that already exist. Training and tutorials will also be offered.

The stakeholder group with four senior academics in the field of imaging will benefit from software and my computational knowledge. This will feed into the Bragg Centre, the Astbury Centre and other imaging facilities on campus. I will also mentor one of their key researchers who will support other researchers. I will also contribute to the N8 centre of excellence so that any new HPC systems will be suitable for imaging.

Computational skills will be upgraded on campus through the computational networks I lead and that will be advertised across the N8. Through business development links at Leeds City Councils I will aim to find commercial speakers and network members. There will be twelve meetings each year across the three networks and these will include a one-day conference for each network.

The general public will benefit from visualizations and artists commission each year. These will be used broadly at science festivals eg, Leeds Science Festival, Astbury Conversations, Light Night Leeds, Otley Courthouse as well as at events at DLS and SuperSTEM.