Mesh Comparison - Day 3

DSME-(Histogram-based) Qualitative Mesh Comparison in between ship CAD systems.

%-------------------------------------------------------------------------------

Goal : Develop the module for 3D mesh comparison, in ACIS environment

Todo :

1. Study

• Qualitative shape (search, detection, matching, comparison, ...)
• MeshLab(VCG) or OpenMesh library
• ACIS & HOOPS

    2. Implementation

• Meshing(Solid, B-rep to facet data)
• Sampling on mesh surface

%-------------------------------------------------------------------------------

• ACIS & HOOPS

- ACIS : Geometric modeling kernel, developed by Spatial.

- HOOPS : Visualization software package that can be integrated with ACIS.

• Meshing(Solid, B-rep to facet data)

- On top of the Hanra's implementation, I developed the module that converts existing geometry in ACIS into vertices and face data.

- For the validation, module for writing the data into the .obj format also developed.

- The quality of the faceting can be differ according to the several parameters in Refinement.

(Top) Source model in ACIS, visualized in HOOPS
(Bottom) Converted mesh model in obj format, visualized in MeshLab(27095 vertices)

• Sampling on mesh surface

- For the mesh comparison based on "Shape distributions". points on the mesh surface should be done.

- To calculate the appropriate distributions that can catch the shape characteristics of geometry, points should be sampled (evenly?) on the mesh surface.

- I developed the module with the method described in HERE.

 

- Sampled points are recorded by xyz format, to validate using MeshLab.

(Top) Source model, 27,095 vertices.
(Bottom) 10,000 sampled points.

%------------------------------------------------------------------------------

Additional Resources

Mesh Comparison - Day 2

DSME-(Histogram-based) Qualitative Mesh Comparison in between ship CAD systems.

%-------------------------------------------------------------------------------

Goal : Develop the module for 3D mesh comparison, in ACIS environment

Todo :

1. Study

• Qualitative shape (search, detection, matching, comparison, ...)
• MeshLab(VCG) or OpenMesh library
• ACIS

    2. Implementation

• TBA...

%-------------------------------------------------------------------------------

• Qualitative shape (search, detection, matching, comparison, ...)

- Method will be based on THIS, "Shape distributions".

- It looks good for the first step implementation.

- Unfortunately, it seems that there is no public library that can be easily integrated into ACIS environment.

- As far as I know, for the development of the "Shape distributions" approach, three major calculation step should be implemented

1. Sampling - (over)sample the vertices form the surface, for distribution calculation.
2. Distribution calculation - A3, D1, D2, D3 or D4 metric calculation among the sampled vertices and counts it to generate the histogram.
3. Distribution comparison - Turn histogram into distribution function and calculate the dissimilarity.

• MeshLab(VCG) or OpenMesh library

- Not going to use the VCG or OpenMesh library, since it contains much more functionality than we needed.

- It means the entire system will be heavy if we use that.

- Instead, the segments of codes or applications will be developed or integrated in ACIS.

- At this time, the resource from HERE is tested for the sampling task.

Left picture shows original model with 298 vertices for 492 faces.
Right picture shows 10000 sampled vertices.
The model is from the Princeton Shape Benchmark

%------------------------------------------------------------------------------

Additional Resources