Structurally-Aware Mesh Simplification Using Skeletonization and Deep Learning (COMP4610/COMP8610-2024) - README
Group 40
- Ananya Sharma (u7669839)
- Manindra de Mel (u7156805)
- Rohit Minstry (U7034818)
This project explores two novel methodologies to enhance the traditional Quadratic Error Metric (QEM) approach for mesh simplification. The integration of skeletonization and deep learning techniques aims to preserve both geometric fidelity and intrinsic structural features of 3D meshes.
We would like to credit the original QEM Mesh simplification author
├───data
│ ├───CNN
│ │ └───output
│ ├───models
│ ├───output
│ └───Skeletonization
│ └───output
├───docs
└───util
├───CNN
└───Skeletonization
- data/CNN/output/: Contains the output files generated from the CNN-based mesh simplification.
- data/models/: Contains saved models and related files.
- data/output/: Contains the output files generated after mesh simplification.
- data/Skeletonization/output/: Contains the output files generated from the skeletonization process.
- docs/: Documentation files related to the project.
- util/: Utility scripts and modules used in the project.
- util/CNN/: Scripts related to the CNN-based mesh simplification.
- util/Skeletonization/: Scripts related to the skeletonization process.
To run the mesh simplification code you need to run simplification.py
-i, --input
: Input file name (required)-v
: Target vertex number-p
: Rate of simplification (ignored if-v
is specified)-ix, --important_indices
: File with important vertex indices
(With Skeletonization)
python simplification.py -i data/mesh.obj -v 100 -ix data/Skeletonization/output/important_indices.txt
(Without Skeletonization)
python simplification.py -i data/mesh.obj -v 100
To generate a text file containing the important vertices and edges, run the following script:
python util/Skeletonization/main.py
The code for skeletonization is inspired from the work done by author
This command runs the mesh simplification process on data/mesh.obj
, targeting 100 vertices, and taking into account important vertex indices from the specified file.