images.imutil.imstat fits a parabola to determine MODE??

Guest wrote on Sep 12, 2006

Hello,

Please help me be sure that I understand how the mode
is derived by the images.imutil.imstat task.

My read of the help file makes me think that after
the final round of clipping, a histogram is produced
and a parabolic curve is fit to the histogram within the range
of bins between the user specified "lower" and "upper"
data limits and with the maximum bin being used as the starting point for
the center/peak of the parabola. After the fit residuals are minimized
an interpolation between
histogram bin centers is performed to establish the precise pixel value
related to the peak of the fitted parabola.

If that is not correct, please clarify what is done as
I'm not sure how else to interpet the help file's description:
"The mode is computed by locating the maximum
of the data histogram and fitting the peak by parabolic
interpolation."

Thanks,
Ed Smith

Post generated using Mail2Forum (http://www.mail2forum.com)

---

Mike Fitzpatrick wrote on Sep 12, 2006

Ed,

I believe you have it right: The mode is the peak of a parabolic fit to the high point of the histogram bins and not simply the highest bin.

Cheers,
-Mike

---

Ed Smith wrote on Sep 12, 2006

Thanks Mike! I was just having trouble wrapping my brain
around "parabolic interpolation". -Ed

---

Last post on Sep 12, 2006