As a mesh reconstruction technique from point cloud data, I have experimented with famously known “Marching Cubes” algorithm [1]. Briefly marching cubes reconstructs a point cloud data by discretizing the data domain with spatial data structures such as grids or voxels then by an implicitly defined signed distance function it determines which voxels contain the queried point. After finding the containing cell, according to the location of point to the corners of containing cell, a shape is approximated per cell. Repeating this for all containing cells, gives us a reconstructed mesh. Details about marching cubes can be found in [1] in detail. My experimentation results are as follows:

As you can see from the images above, I have a small bug in my implementation which I am working on it, I will update this post whenever I find it. Hope you liked it $\triangle$

References:

1 - http://paulbourke.net/geometry/polygonise/