I used NIH Image some years ago for this purpose. ImageJ was based on that package.
The important thing for quantitative densitometry is to get a good set of OD
filters or a step filter. You can order these from Edmund Optical, Oriel or
Newport. You plop these down on your flatbed scanner with a transparency
attachment and scan them (you need the transparency attachment, otherwise you'll
be looking at reflected light). You then get the "value" for the pixel in
grayscale mode (for 8bit that's 256 shades of gray). You take several of these
measurements for each OD and plot a regression curve. Congrats, you now have a
Take whatever you want to scan (film, SDS-PAGE, etc) and place it on the scanner
and collect a scan. You can use ImageJ to give you a plot of values versus
position along a line (perhaps a lane on a gel). Using your regression formula
from your standards you can calculate the ODs for such a plot.
Matt Rand, who may be in your department, has extensive experience with this.
Quoting "Garret D. Langlois" <[log in to unmask]>:
> This might be a long shot, but do any of the biology/computer science hybrid
> people have experience using ImageJ for densitometry? Your welcome to e-mail
> back direct if you do. Thanks.
> Garret D. Langlois
> (802)656-9785, HSRF 415A
> Dept. of Anatomy & Neurobiology
> 149 Beaumont Ave., Burlington VT 05405
> College of Medicine, University of Vermont