Research Associate / Senior Research Associate positions available

In the framework of the EPSRC funded AssenSAR project, we have available two positions focussing on the development of novel algorithms and associated software tools for the mathematical and statistical characterization of the sea surface and on the solving of a range of inverse problems involving SAR imagery, respectively. For more details see

GUI for granule segmentation in TEM images

Link added to Matlab code for the algorithm proposed in: David Nam, Judith Mantell, David Bull, Paul Verkade, Alin Achim, A Novel Framework for Segmentation of Secretory Granules in Electron Micrographs, Medical Image Analysis, Volume 18, Issue 2, February 2014, pp. 411-424.

See Software section for details.

Bayesian MAE image denoising code made available

13 years after the publication of my IEEE TMI paper on Bayesian denoising of ultrasound images with alpha-stable distributions I still get the odd request for the code that implements that algorithm. I have today finally decided that it is more efficient to have it up on a website than keep e-mailing it … 😐 You can find it in the ‘Software‘ section of the website, while below are some explanations, copied and pasted from the e-mail I kept copying and pasting for all these years:

Dear …,

Thank you for your interest in our work.
I’m sending attached some of my code which will hopefully help you re-implement the method in that paper. Please note however that not all the attached code is the one used to derive the results therein as I lost some of the original files during my trips from one institution to another across Europe before getting my current job. The main differences are:

  • In order to run the code, you’ll need to go to and download the matlab implementation of the dual-tree DWT. My original implementation used the separable DWT embedded into the cycle-spinning algorithm.
  • In the present form, my code does additive noise removal. If you want to remove speckle you’ll need to use a log transformation in the beginning and an exponential one at the end.
  • In the main function, line 67 reads im = stablepdf(x1,alpha,0,disp,0,2); I cannot provide you the code for calculating stable density functions as at the time it was provided to me by John Nolan under promise that I won’t recirculate it. You would thus need to go to and buy the STABLE code … (or write your own function to compute stable densities)

Hope this helps!