Re: using a scanner as a densitometer?

Tom,

you are absolutely right about the learning curve it's very steep especially
now with all these packages available but but but to use a series of 21
values or up to 5 such series and plot a (intelligent) smoothed curve for
each is kind of child play for those who have even a little experience with
R. I already have this kind of function and I would gladly share it with
others after a bit of finishing. I'm sure something very similar can be done
with Excel which most of you could have already. I'm sure many alternative
are availlable and each present some level of learning even using a piece of
paper, a pen, etc to draw your own graph manually present some learning for
those who never did this before. For the limited number of times these graph
have to be done, I wonder if the paper solution wouldn't be the most user
friendly after all.

Regards
Yves

----- Original Message -----
From: "T.E. Andersen" <postlister@microscopica.com>
To: <alt-photo-process-L@usask.ca>
Sent: Wednesday, February 22, 2006 8:33 AM
Subject: Re: using a scanner as a densitometer?

> Hi all,
>
> R is great, but the learning curve is very steep, esp. for
> non-programmers (like myself). I have used it a bit, and have colleagues
> who swear by it. I've never found the time to learn it well enough for
> routine use, though.
>
> I would like to suggest an alternative that may suit more people: ImageJ
> (may be downloaded freely from http://rsb.info.nih.gov/ij/index.html)
>
> Many scientists use ImageJ to evaluate electrophoresis gels. This is
> basically an application of densitometry! Indeed, many of the image
> analysis options in ImageJ has a component related to densitometry.
>
> ImageJ is simple to use.
>
> I think doing a simple line profile along a step tablet should be
> adequate for most applications. ImageJ would have to be calibrated (see
> http://rsb.info.nih.gov/ij/docs/examples/calibration/index.html). When
> that's done, all you have to do is to draw a line (or a rectangle for
> better averaging) along the step tablet, and press ctrl-K to plot the
> profile. Actual values may be read right off the y-axis, or by using the
> list function in the plot window.
>
> This approach may not be entirely accurate, but I would be surprised if
> the inaccuracy is large enough to cause problems in
> alt-photo-applications. If better accuracy is needed, I think a real
> densitometer is the way to go.
>
>
> Best regards,
> Tom Einar Andersen
Received on Wed Feb 22 08:21:14 2006