I needed a way to visualize a subset of a 2D variable derived from a 2D dataset, _together_ with the original data and in its original coordinates. Since a plain 3D plot decouples all this info (and becomes illegible for data containing very noisy regions) I decided to show the calculated 2D descriptor as hue variation on the original 2D data (represented as gray level).
It is a suggestive way of visualizing data and especially effective when the 2D subset is a lot smaller than the original 2D data (for instance when the variable to plot is uninteresting for the most part and it would only skew the z-scaling).
Following the discussions on the FileExchange and Newsgroup it seemed to me that more people might benefit from this, so I decided to post it.
>> imgRGB = hotplot(BACKGND, FOREGND, POSITION, 1);
when FOREGND is an N-length list of values and POSITION is an N-by-2 list of coordinates (in the BACKGND referential) ;
>> imgRGB = hotplot(BACKGND, FOREGND, MASK, 1);
when BACKGND and FOREGND are both [M x N] and MASK is a boolean, obtained for example as
>> MASK = FOREGND > SomeThreshold;
or run testHotPlot.m to visualise a sample data set; the physical meaning of this dataset is :
FOREGND - neural impulse energy
BACKGND - average neural potential
for ideas, suggestions, etc. mail me at
tudima at y a h o o dot com
v.02 : more flexible calls (it now accepts a full FOREGND), improved input handling