Publication:

Proceedings of Pacific Graphics 2014

Authors:

Yixin Zhuang^1,2, Ming Zou², Nathan Carr³, Tao Ju²

Affiliations:

¹National University of Defense Technology (China), ²Washington University in St. Louis (USA), ³Adobe

Shortest paths, colored by their lengths (blue: short; red: long), in the Euclidean metric (a) and our anisotropic metric (b) from a single vertex (red) in the Fertility model, and a live-wire network where each wire (black) is a geodesic in our metric between two seeds (blue) (c).



Ridges and valleys (a), automatic segmentations (b), curves sketched manually by an artist (c), and a segmentation created using our live-wire tool guided by the artist's sketch (d). Arrows in (d) point to wires defined by scribbles. Note the difference between algorithmic curves and human-drawn curves, and the sparsity of nodes in our curve network.



Left: Comparing geodesics in the Euclidean (red), Campen's (magenta), Pottmann-Kovacs (green) and our (blue) metrics when endpoints lie on a curved ridge (top) and a curved valley (bottom). Note that Pottmann-Kovacs geodesics tend to avoid the features when flat regions are nearby. Right: More segmentation results.

Abstract

We present an interactive method for mesh segmentation that is inspired by the classical live-wire interaction for image segmentation. The core contribution of the work is the definition and computation of wires on surfaces that are likely to lie at segment boundaries. We define wires as geodesics in a new tensor-based anisotropic metric, which improves upon previous metrics in stability and feature-awareness. We further introduce a simple but effective mesh embedding approach that allows geodesic paths in an anisotropic path to be computed efficiently using existing algorithms designed for Euclidean geodesics. Our tool is particularly suited for delineating segmentation boundaries that are aligned with features or curvature directions, and we demonstrate its use in creating artist-guided segmentations.

Video

Downloads

Acknowledgements

We thank Daichi Ito for providing the manual segmentations in Figure 11, and authors of papers [YBS05, CSAD04, dGGDV11] for providing code and/or data. This work is supported in part by NSF grants (IIS-0846072, IIS-1302200, IIS-1319573), an NSF China grant (61379103), and a gift from Adobe.

BibTeX

@inproceedings{zhuang2014anisotropic,
title={Anisotropic geodesics for live-wire mesh segmentation},
author={Zhuang, Yixin and Zou, Ming and Carr, Nathan and Ju, Tao},
booktitle={Computer Graphics Forum},
volume={33},
number={7},
pages={111--120},
year={2014},
organization={Wiley Online Library}
}