Description:
Finally, we reached the last step of the procedure, the most important one, because here we can see if there are any errors in the previous steps, what went wrong and if the method used could be improved.
The goal of this step is to approximate the volume with voxels (tiny cubes) and then generate a mesh out of this cloud of voxels. For this we used an in-house program that does the volume carving and the volume editing and then a free software, the Visualization Toolkit in order to do the marching cubes.
Code:
We did not write any code at this point.
Different smoothing factors and isosurface values were tried. Here are three of the results, at a resolution of 256x256x256:
Comments:
That was the most interesting part of the project. Here I found out that the calibration is not perfect, the data is not perfect, the alignment is not perfect, so, that is why the final version looks how it looks.
Anyway, that was done at a resolution of 256x256x256 voxels. I am sure that increasing the resolution to 1024x1024x1024 would give a lot better details. The only problem with that is the disk space and the processing time, which would be multiplied by a factor of 64 and if now we are using about 32Mb for the 256 resolution, the math speaks for the requirements of this high resolution.
We learnt that the method is very good, provided that high quality equipment is used to eliminate the errors, a lot of disk space is available and a lot of free time in order to carefully acquire the images and process them.
