Symmetry-Aware Mesh Segmentation into Uniform Overlapping Patches

This companion website supports our paper Symmetry-Aware Mesh Segmentation into Uniform Overlapping Patches.

Supplementary Material

In this document (PDF, 4 MB), we provide supplementary material with a proof to bound the asymptotic number of patch neighbors, as well as some additional results on mesh segmentation and patch growing experiments.

Videos

In this video (MP4, 655 KB), we show the iterative sampling of 500 patches for uniform mesh segmentation of the Happy Buddha dataset.

In this video (MP4, 749 KB), we show the agglomeration of 100 patches with an overlap ratio increasing up to 0.5 for growing the segmentation of the Bunny dataset.

In this video (MP4, 640 KB), we show the iterative sampling of 200 patches with no seeds for asymmetric uniform mesh segmentation of a scanned face from the Basel Face Model database.
In this video (MP4, 635 KB), we show the iterative sampling of 200 patches with no seeds for symmetric uniform mesh segmentation of a scanned face from the Basel Face Model database.
In this video (MP4, 718 KB), we show the iterative sampling of 200 patches with Farkas seeds for asymmetric uniform mesh segmentation of a scanned face from the Basel Face Model database.
In this video (MP4, 720 KB), we show the iterative sampling of 200 patches with Farkas seeds for symmetric uniform mesh segmentation of a scanned face from the Basel Face Model database.

In this video (MP4, 350 KB), we show the agglomeration of 100 patches with an overlap ratio increasing up to 0.5 for growing the asymmetric segmentation of the Dino dataset.
In this video (MP4, 332 KB), we show the agglomeration of 100 patches with an overlap ratio increasing up to 0.5 for growing the symmetric segmentation of the Dino dataset.

Code

In this archive (ZIP, 32.1 MB), we provide an implementation of our proposed methods in the form of MATLAB source code to reproduce the experimental results obtained in the paper.

For the moment, the archive contains the main methods to perform asymmetric and symmetric uniform mesh segmentation and patch growing as well as texture stitching. More functions for texture synthesis will be added as we clean the code and obtain authorization to distribute or find relevant ways to link to licensed code and data that we used. We will also move the code to github soon.

Valid HTML 4.01 Transitional