Clusterring of 3D point cloud

Hi,

I am having 3D data to fit surface on. I need to partition data to use different degrees while surface fitting. Any suggestion will be highly appreciated.

[darova]

You can use simple code

ind1 = xmin < x & x < xmax;
ind2 = ymin < y & y < ymax;
ind = ind1 & ind2;                % indices of point you want

[Srinu Vasu]

Dear darova,

Thanks for the answer but the data im having is non uniform and complexity will be varying over the span of the data. First i tried above method but it is not useful for complex data. I'm attaching the figure and 3D data (This is 3D data of face) for your reference.

[darova]

Here is what i achieved using griddata()

Are you trying to get equation of this surface? What you mean by "fit"?

[Image Analyst]

What do you mean by "surface fitting"? Do you mean like fitting a plane through all the points? Do you mean something like a Delaunay surface to contain all the points?

[Srinu Vasu]

Surface fit what i mean is simply i want clear surface on this 3D point data.

When i'm using curve fitting app for surface for this 3D data the output i'm getting is not satisfactory( The person in the face should be identified). So i want to devide the 3D point cloud(e.g nose ,forehead,cheeks) to furhter use different degrees of surface to fit.

This data is not scanned data. It is reconstructed data from photogrammetry.

[darova]

I see the face