Comments and Ratings

You have introduced a lot of unnecessary conversions in your code for converting from string to vector:
This line suffices for doing the conversion:
binVec = uint8(binStr) - uint8('0');
If you want a double vector, you can just do a double(binVec).

I'll admit I'm not entirely sure what purpose this serves as a Matlab utility. Since it seems to be of mainly educational value, why not add some links to more information about BMI? Perhaps these are a start:

http://www.cdc.gov/nccdphp/dnpa/bmi/

http://www.cardiometabolic-risk.org/

While I won't get into the questions of whether BMI is an accurate indicator of health, nor will I get into the issue of mono-numerosis, BMI surely is one indicator of health. So the author might have added some indication of what a reasonable BMI should be. Otherwise, its just a random number. The plot might have been done so that the line varies in color, scaled so that good BMI numbers would be green, turning to red at the high end.

The help itself for bmi.m is good. Its readable and concise, defining the inputs and outputs. I'm not terribly sure why the units on the inputs are in pounds and inches, yet the returned BMI has SI units. Tradition?

I found an H1 line, but no error checking on the arguments. OTOH, the arguments are pretty simple, so I'll excuse that lack. The code is pretty simple, so nothing to go wrong, and I did find internal comments before many lines.

Now, if I can only figure out how to be nine feet tall, I'll have a perfect BMI myself. ;-)

Rick,
this code can be easily rewritten to get rid of the imageprocessing toolbox dependency.

Also, I am pretty sure you want to take the following line out of your for-loop.
C = zeros(N,3);

Without the appropriate comments, it is difficult to tell but why use the line below line?
gr = gr(:,1);