Skip to main navigationSkip to main content
The University of Southampton

Research project: Filling in the gaps: compressed sensing for x-ray computed tomography

Currently Active: 

X-ray computed tomography (XCT) is a technique that uses x-rays to image the three dimensional structure of objects. XCT is routinely used for medical diagnostics, is a powerful tool in many scientific areas, from archeology to biology and is becoming increasingly common as a tool in manufacturing, where it can be used to inspect the internal integrity of components and is able to provide dimensional measurements of parts with complex shapes and internal structures.

Compressed Sensing is a novel and powerful mathematical framework that enables us to recover signals and images that are only partially observed. For example, in XCT, three-dimensional images are constructed from a set of x-ray projections, taken at several angels around the object. If limited projections are available or if access is restricted to certain directions, because the object is too large, traditional XCT reconstruction methods do not work and advanced methods, such as compressed sensing techniques can be used.

Exploiting the significant advances made in the field of compressed sensing recently, we are working on novel methods to solve the XCT reconstruction problem in several challenging imaging settings, including restricted view and low number of projections imaging as well as short exposure time, low contrast and high noise settings.


These methods are helping our industrial partners to build up additional advanced capabilities in non-destructive testing and dimensional measurement, thus further increasing safety, reliability and sustainability of their manufacturing processes and products whilst reducing costs.

Project Partners

This work is done in close collaboration with several industrial partners such as the National Physics Laboratory (NPL),
Nikon Metrology Ltd and
as well as the
Collaborative Computational Project in Tomographic Imaging (CCPi) and the University of Southampton's Multidisciplinary, Multiscale, Microtomographic Volume Imaging lab (μ-Vis)


This work is funded by the Collaborative Computational Project in Tomographic Imaging (CCPi) and through the EPSRC's visiting fellowship scheme.

Related research groups

Signal Processing, Audio and Hearing Group
Shepp-Logan vhantom
Shepp-Logan vhantom
Sinogram of 9 projections
Sinogram of 9 projections
Total Variation
Total Variation
Re-weighted Total Variation
Re-weighted Total Variation
Share this research project Share this on Facebook Share this on Twitter Share this on Weibo
Privacy Settings