Skip to content

Friday Whatsit for August 13, 2010

August 13, 2010

What is it?

Everyone is getting so good at the Friday Whatsits I thought I would put up something a little more challenging.

Imaging is not just about taking pretty pictures, although reviewers of manuscripts and grants often want just that: a pretty picture of whatever it is a scientist is trying to describe in words.   Pictures do serve to illustrate data, but they can also be used to generate data, if you can find a way to extract the information you need.  The image on the right used to be a picture of some round little cells sitting on a coverslip.  Until I transformed it into a useful tool by manipulating it with the computer.  Just the picture was transformed, mind you.  Innocent cells were not harmed in the making of this image.

What it the name of this tool and what information does it give?

Possible hint for the geometrically minded:  Ignore the colors but know this, each colored band is one pixel wide and in the non-color coded (greyscale) version, the centers were all intensity 0, the next band is 1, the next is 2, etc until it bumps into another band.  The whole scheme can also be inverted.

Muhahaha. . .  challenging, but much closer to real science than the last few weeks.  More sciency, as one of my readers suggested.  Go forth and guess!  Answer on Monday!

From → Friday Whatsit

  1. April permalink

    OOh is it a fourier transform??

    • Mmmm. . . it’s not a Fourier transform, although I thought about posting one of those for you and Keith. Really, I did. That would blow some minds! Unfortunately even though I know what an FT is and what you can do with it, I can’t explain what it is without it sounding like gibberish. The equations make my brain go all to goo.

      You’re close though! I’m sure you’ll get it if you think about it. You might have even used one. If your objects are larger than infinitely small points (and well sampled) you can make one of these.

  2. April permalink

    All my objects are infinitely small points…
    let’s see we use point spread functions, Fourier transforms, something with fractals might be cool. We use bias frames, dark frames, flat. To me it really just looks like it has halos around it.

    • I know all your objects are infinitely small points, which is why you probably never use a Euclidean distance map : ) Plus your pixel size is on the order of light years and mine is nanometers, so you would have way more error in your measurements than we do.

Comments are closed.

%d bloggers like this: