Posted also here, where it's been put on hold. I did modify it afterwards though.


I'm an academic researcher in the US working in an area called "statistical shape analysis", which is intimately connected with medical image computing. It deals with how to mathematically compare two images, or predict (regression problem) how the future images (say MRI image of Alzheimer's patients, or images of protein structures) will be.

Let's say I'm working on statistical modeling of hippocampus (a part of the brain) to infer about which stage the Alzheimer's disease is. So I obtain the segmented MRI images, perform statistics on it to predict future segmented images. But I can't tell how exactly the image itself will look like, but only how the segmentation will be like. I hand this prediction to a doctor. How will that be helpful to a doctor who's treating Alzheimer's? Of course, intuitively it's clear it's helpful, but I'm asking what exactly the doctor will do with these kind of computational data.

Also, could you mention some hospitals in Europe where they use informaticians/statisticians to collaborate with doctors for the above kind of work?

  • $\begingroup$ I don't know how many doctors you are going to find on a DSP board, but it's probably not very many. $\endgroup$ – AnonSubmitter85 Mar 8 '16 at 0:58
  • $\begingroup$ Well, it seems to me that you yourself don't know how your research interrest can be of use - which is sad, because every research field comes with a vision of the future and how it could be beneficial. However, since you do have interesting data and predictions, here is something that came very clear to me: The simpler the technique and the simpler the results are understandable, (sadly) the more likely it is to be of use in a clinical field. So, try to crunch your results down to a few understandable numbers, e.g. predicted volume change of a VOI or something like that. $\endgroup$ – M529 Mar 8 '16 at 19:40
  • 1
    $\begingroup$ Most papers in computer vision and image processing is focus on the pattern recognition of some disease, in other hand there are many papers based on diagnostic task, where the image is "enhanced" to give to the doctor something like a "hint" to prepare your diagnosis. $\endgroup$ – Darleison Rodrigues Mar 8 '16 at 23:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.