When you need a CT scan ask for this | Hacker News Reader

When you need a CT scan ask for this

freepress.intel.com

11 pointsMitt7 comments

The answer to the question Baker posed lay in a set of mathematical rules called an algorithm.

I never cease to be amazed just how low the math level is in our culture. And how acceptable it is to be publicly that math ignorant.

It was difficult, for example, to change the algorithm, which worked best in a single-threaded environment, into one that could take advantage of multi-core processors.

Aha, so the parallelized it!

"We have been able to reduce X-ray doses to previously unthinkable levels," said Professor Johan de Mey, head of the radiology department of University Hospital in Brussels, Belgium. That is opening up the benefits of CT scans to a wider variety of patients.

Great!

The terrible way in which this article explains what was done, is doubly ironic considering how it is an article about math and programming literally saving lives.

Using the same phrases you consider "terrible", the writer was able to make this piece accessible to a wider audience. I would suggest that articles like this, which describe technological breakthroughs in language the lay public would understand, are good for the science and engineering communities.

By the way, the article isn't about "math and programming saving lives", it is about an advancement in technology that reduces the amount of radiation patients are exposed to (in CT scans) so that there's a lower risk of radiation poisoning.

My advisors work is in this - part of it is making better use of the statistical properties of the noise in the reconstruction of the image.

Stochastic Resonance

http://en.wikipedia.org/wiki/Stochastic_resonance

That was not what we were doing - It was more like weighting measurements proportional to their statistical variance, so that you penalize noisier measurements. In this process, you typically apply some sort of regularization to the objective function so that it converges to a smooth solution. There are some interactions between the weights chosen and standard regularization that creates non-uniform spatial resolution propeties, so much of the work was spent designing roughness penalties that played nice with the weightings.

Hmm. Pretty much the inverse of SM.

http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=5358...

Reply on news.ycombinator.com