Abstract:

... that form ever follows function. This is the law.  
                        — L. Sullivan (1896)

Extracting semantically related parts across models remains challenging, especially without supervision. The common approach is to co-analyze a model collection, while assuming the existence of descriptive geometric features that can directly identify related parts. In the presence of large shape variations, common geometric features, however, are no longer sufficiently descriptive. In this paper, we explore an indirect top-down approach, where instead of part geometry, part arrangements extracted from each model are compared. The key observation is that while a direct comparison of part geometry can be ambiguous, part arrangements, being higher level structures, remain consistent, and hence can be used to discover latent commonalities among semantically related shapes. We show that our indirect analysis leads to the detection of recurring arrangements of parts, which are otherwise difficult to discover in a direct unsupervised setting. We evaluate our algorithm on ground truth datasets and report advantages over geometric similarity-based bottom-up co-segmentation algorithms.

Results:

We present a top-down indirect analysis to discover corresponding parts (bottom) across objects, even with significant geometric and topological differences (top).

Consistent part arrangements extracted using our unsupervised algorithm. For each model collection, we show only a subset of the results showing both the extracted part arrangements and parts. Unclaimed components are in gray.

Different representative part arrangements extracted for various model collections. Note that our unsupervised analysis relies only on available geometry information and has no access to part names or additional tagging information. The symbols k and N denote the number of part clusters and the number of models in the shape collections, respectively.

@article{zcam_partArrangement_eg14,
AUTHOR = "Youyi Zheng and Daniel Cohen-Or and Melinos Averkiou and Niloy J. Mitra",
TITLE = "Recurring Part Arrangements in Shape Collections",
JOURNAL = "Computer Graphics Forum (Special issue of Eurographics 2014)",
YEAR = "2014", 
numpages = {10},
}