On Mon, 2 Jul 2007, Malcolm J. Currie wrote:
>> The KAPPA HISTAT application uses a data histogram to calculate the median
>> of a data set. It says that it also calculates the mode but it actually
>> seems to calculate a fudge factor and in my data is a bad value. Is there
>> a reason why it can't simply write out the peak of the histogram?
>
> There are two new histogram-based methods (see new parameter METHOD) in
> addition to the Pearson formula. For a selection of test data these
> work pretty well, although the mars_3pos image the theory-based
> presciption for the optimum number of bins looks in adequate as there is
> some fine detail there, although it may be through the addition of
> exactly zero-valued pixels.
Thanks Malcolm. Andy Gibb will try this out this week.
>
> I have toyed with the idea of locating the peak first, then selecting a
> narrower range about that based upon the interquartile range, and then
> recomputing the bin size using that focussed histogram. However, I
> don't want to make it over complex and slower given Tim's 10Hz
> requirement. Give it a try and let me know if the new options are
> adequate.
>
Ok. We'll let you know how much time we've got available (currently we are
doing the per-frame processing in 0.4sec and group processing every 10th
in 1.2 seconds which is fast enough not to drop a frame.
--
Tim Jenness
JAC software
http://www.jach.hawaii.edu/~timj
|