Inquiry Regarding Discrepancies in Mesh Comparison Results Between CloudCompare and Geomagic Control X

[MazenMeso]

First and foremost, I would like to express my gratitude for the substantial investment in developing and maintaining the exceptional software you provide, as well as the continuous support offered through your platform.

I am currently engaged in research for my thesis, focusing on Geomagic Control X software. However, due to its proprietary nature, I have sought alternative solutions and thankfully came across CloudCompare, which has proven to be a valuable resource.

To delve into the intricacies of these software tools, I conducted a mesh superimposition using CloudCompare—employing point pairs alignment followed by fine registration. Subsequently, I computed the cloud/mesh distance, resulting in the following values: Min = -0.686 mm, Max = 0.765 mm, Mean = 0.023357 mm, and RMS = 0.135114 mm.

To maintain consistency, I exported the aligned meshes and imported them into Geomagic Control X without initiating any additional alignment steps. I performed a comparative analysis on the entire meshes (using 3d compare tool), yielding values of Min = -0.6661 mm, Max = 0.6574 mm, Mean = 0.0175 mm, and RMS = 0.1276 mm—values that exhibit noticeable discrepancies from those obtained with CloudCompare.

My understanding, based on my research, is that both software tools compute similar analyses to derive RMS values (as outlined in the link: [https://support.3dsystems.com/s/article ... uage=en_US]). However, the observed disparities prompt my inquiry into the reasons behind these differences. I acknowledge the variance in the alignment algorithms employed by the two software tools.

I also employed MeshLab software, which provided results consistent with those obtained through CloudCompare, but again, discrepancies were noted when compared to Geomagic Control X.

Your insights into the underlying factors contributing to these variations are crucial for my thesis. Understanding these differences will help justify the transition to using CloudCompare in my research.

[daniel]

Hi,

Are you able to export per-point distances with Geomagic?

[MazenMeso]

yes
(refer the attachments)

You can review the calculation methods used in Geomagic by following this link.
https://support.3dsystems.com/s/article ... uage=en_US

[daniel]

Si I saw the link, but it's not super clear how they manage the cases when the projection of the point on a triangle falls outside of the triangle (and potentially outside of any triangle). Similarly to MeshLab, we use the algorithm defined in the old 'Metro' tool (https://github.com/cnr-isti-vclab/vcgli ... apps/metro) or alternatively by 'Geometric tools' (see https://www.geometrictools.com/Document ... angle3.pdf). Basically, we use the distance to the closest edge or the closest vertex.

If geomagic is following a different approach, this may explain the differences.

Can you export the two clouds (with their respective distances) and share them with me? (admin@cloudcompare.org). Or you can also load both in CC, interpolate the scalar field of one onto the other (make sure they have different names), and then compare the 2 scalar fields with the 'SF Arithmetic' tool. This may help us to visualise the differences, and if they are in specific parts of the mesh.