Interpolate 3d points with griddata or scatteredInterpolant - What's v?

Hello,

I have a question that has been asked a few times on different ways but I have not been able to understand it.

My data:

I have a tooth as in the upload, which is the result of a mean calculation from other teeth. As you can see in the picture, it is fragmented as a result from the calculation.

I would like to make an interpolation, so that these fractures are filled.

As I have understood, I should use griddata or scatteredInterpolant, both having a variable v which I don't understand. Can someone explain to me or show me in the right direction, how can I understand this variable v, so that next time, when I have to interpolate something (which will happen), I can do it on my own?

Of course I would be very happy if someone gave me the solution for this explicit example, but I honestly want to understand it.

Thank you :)

[Bruno Luong]

"As I have understood, I should use griddata or scatteredInterpolant"

No you have not have a function to interpolate.

Here you have a scattered points that supposes to represent an object surface that you want to fill at the place where it is not dense.

Both are different problems.

Therefore griddata or scatteredInterpolant are wrong tools IMO, and it's normal that you don't understand the role of V, there is none in your problem.

[Diego Hens]

Ok, understood. Thank you.

And would you happen to know what I can use instead?

[Bruno Luong]

You need first to reconstruct the mesh from point cloud (search those keywords).

https://www.mathworks.com/matlabcentral/fileexchange/63730-surface-reconstruction-from-scattered-points-cloud

https://www.mathworks.com/matlabcentral/fileexchange/63731-surface-reconstruction-from-scattered-points-cloud-open-surfaces

I though you start with STL as I follow some of your question. It's pitty that you throw away the connectivity and need them back by using as postprocessing one of such tool.

Such tools might work but the performance varies.

Then you need tool such as remeshing, there is a great SW called MESHLAB.

MATLAB is poor in term of mesh processing.

[Diego Hens]

Thank you very much. I will look into it soon and then come back with my results.

I started with STL, yes. But after a few steps, the connectivity is of no use anymore, since the points don't correspond anymore.

[Diego Hens]

Thanks, this will be very useful!