Comments and Ratings

Send the data on my e-mail I need to see them.

Warning: MEX could not find the library "glut"
         specified with -l option on the path specified
         with the -L option.
BatchFaceColorGradImpl.cpp
OffscreenGL.h(12) : fatal error C1083: Cannot open include file: 'GL/glew.h': No such file or directory
 
  C:\PROGRA~1\MATLAB\R2009A~1\BIN\MEX.PL: Error: Compile of 'BatchFaceColorGradImpl.cpp' failed.

Never used open GL, any help?

You are right, the bug is generated only when the user choose to return distances. This is why I didn't realize before.

Thank you very much

AH AH AH!

That's my latin-english. I'll fix it in the following days.

Thank you for your report

Given:
 
 N=reference +query points

Complexity of the algorithm is linear for random uniformly spaced points. Both for the serach and the tree construction.

 It becomes N^2 (Brute search) for very sparse dataset.

However this patological case can be detected after 5-10% running time. In this case is better to switch to kd-tree data structure. ( this detection is not present in this version)

                    Luigi

Please report questions on my e-mail, this space is dedicated to comments and ratings.

If you have a filled colud there are two main options, depending if your cloud comes from a tetraedral FEM mesh or not.

-If you have tetraedrons, than it is easy just get the boundary triangles.

-If you do not have them, it is more complicated, a ball pivoting or alpha shape tool is needed.

For more info, questions, suggestions contact me on my e-mail.

I like when someone makes things easy, avoiding to complicate the work just to make it seem a good work.

Hi John thanks for the comment,

Honestly I didn't know the ndgrid command. More honestly, now that I know, I don't realize how to use it.

I thought about the 3D arrays as input but I personally prefer the 2D concatenation. Anyway in the next revision I will add this feature, since I thing the assembling indexes stays the same and it wont take a long time to code it.

I am not so confortable of least squares problems, I have just used lsqlin a few times. But I thing that this kind of solvers are slower that the ones for linear system, so the advantages of a vectorization will be minor. The for loop and function call time will be hidden by the solver time. As soon as I have a little free time I will run a few speed test and decide if add this new feature.

Thank you again

Very Very Very Nice

Now is clearer!! what a long discussion did it cause!!!!!

I give this 4 star, one is missing beacuse I thing is a little to simple model. But overall it is nice submission like so many comments testify.

In the end I thing this is the MEX purpose, to create discussion and improve each other.

 Thanks to the Author and all commenters !!

Nice GUI but there is something wrong in the mathematical model.

This not a bouncing ball, in the simplest model, the friction value, I think you mean the friction with the ground, should affect the way ball pivots not length after that the ball stops.

For zero friction the ball just stops why?
-It should instead bouncing without pivoting.

-With intermediate values should slides and pivots dependending on materials data.

-With high friction should pivots without sliding.

If you try to insert a friction value >1 the ball accelerates, and this isn't phisical.

I saw you study mecatronics, you can realize that with Working Model software, or simpler throwing a ball.

I thing it need to be fixed

The Author:
I received an e-mail about problems on Matlab R2007a (7.4.0). The problem is a skinny triangle on mannequin chin. That shoudn't be there.
On my version R2008a everything is working fine the model is recostructed correctly. I would like to know if somebody else got the same or different problems in versions previous to mine. Please send reports to my e-mail

Very useful for binary filesl but do not support ascii format

Wonderful Congratulations,

It is a very usefull routine, with nice graphical plots.

I need to ask you a few question:

How to obtain profile coordinates?

Does it computes the profile for roller follower?

At the end a few possible improvement:

-rocker roller follower instead of translating.
-You can also includes the detachement problem which is very useffull for high frequency cams.

-Please insert a better documentation on how to use use this beatiful utilitie. I took 20 minutes to understand how to start!!!!

Thank you again

From your timing I suspect you did the same error of Tomas Clark

You should turn the profile off, profile only works on m-code, the native convhull is a mex routine, so it only slow down my algorithm.

To be sure please send a report about the way rou ran your test. If there is any problem I'll try to fix it.

Very nice

That'is it Ifound
Mister clark there is nothing wrong in my pc, maybe you should
turn the profile off when you run a new test try this. Especially
before giving low and inappropriate rating.
 
%% PUT THIS IN A SCRIPT AND RUN IT
% profile on
% profile clear
 
%%%%%%%%%%%%%%%%%%%%%%%%
 
%Maybe you forgot this !!!!
 
profile off
 
profile off
 
profile off
 
profile off
 
%%%%%%%%%%%%%%%%%%%%%%%
 
Npoints = logspace(1,6,15);
lengthN = numel(Npoints);
convhull_times = zeros(lengthN,1);
chull2d_times = zeros(lengthN,1);
for i = 1:lengthN
    N=ceil(Npoints(i));
    x=rand(N,1);
    y=rand(N,1);
    tic
    chull1 = convhull(x,y);
    convhull_times(i) = toc;
    tic
    chull=ConvHull2D(x,y);
    chull2d_times(i) = toc;
end
      
figure()
semilogx(Npoints,convhull_times,'g-')
hold on
semilogx(Npoints,chull2d_times,'b-')
legend({'Native convhull';'convhull2d'})
title('Performance comparison')
% profile viewer

Very good code, readable, uderstandable and even fast.
The n-dimensions also make-it flexible.
 
The thing I like the most is a wonderfull example of dot product vectorization with memory management. It helps a lot to understdand how to make m-code more efficient.
 
The only thing i disagree is the comparison with tsearchn.
 I thing its purpose is to find where a point is located inside the hull and not whether he is inside or not. The nan output values are just a secondary objective which occurs in the worst case when a point lies outside.
 
Anyway a five stars, very good work!

1)I don't Know what kind of data you used but I didn't provide any.
The demo file contains just random points generation.
My pc do not support 1e7 due to help memory error, anyway finding the NNG in 1e7 points take much longer than 0.02 sec for both algorithoms.
My algorithom suffers query points outside the bounding box of reference, this should be the case in which ANN is faster.
2)One thing you are right is that ANN is more flexible.
3)From your comment it is not clear what you changed seems just like you copied and pasted the same line.
4)you could at least use your name
5)If you have problems my e-mail is the best place where fix them so please avoid fake comments

Thanks sometimes pdf are better than m-file.

Thank You,
 it was very helpfull, impressive code.

If I can suggest an improvement:
a mex version will be the top of performance.