Unsupervised Cycle-consistent Deformation for Shape Matching
| dc.contributor.author | Groueix, Thibault | en_US |
| dc.contributor.author | Fisher, Matthew | en_US |
| dc.contributor.author | Kim, Vladimir G. | en_US |
| dc.contributor.author | Russel, Bryan C. | en_US |
| dc.contributor.author | Aubry, Mathieu | en_US |
| dc.contributor.editor | Bommes, David and Huang, Hui | en_US |
| dc.date.accessioned | 2019-07-11T06:19:30Z | |
| dc.date.available | 2019-07-11T06:19:30Z | |
| dc.date.issued | 2019 | |
| dc.description.abstract | We propose a self-supervised approach to deep surface deformation. Given a pair of shapes, our algorithm directly predicts a parametric transformation from one shape to the other respecting correspondences. Our insight is to use cycle-consistency to define a notion of good correspondences in groups of objects and use it as a supervisory signal to train our network. Our method combines does not rely on a template, assume near isometric deformations or rely on point-correspondence supervision. We demonstrate the efficacy of our approach by using it to transfer segmentation across shapes. We show, on Shapenet, that our approach is competitive with comparable state-of-the-art methods when annotated training data is readily available, but outperforms them by a large margin in the few-shot segmentation scenario. | en_US |
| dc.description.number | 5 | |
| dc.description.sectionheaders | Synthesis and Learning | |
| dc.description.seriesinformation | Computer Graphics Forum | |
| dc.description.volume | 38 | |
| dc.identifier.doi | 10.1111/cgf.13794 | |
| dc.identifier.issn | 1467-8659 | |
| dc.identifier.pages | 123-133 | |
| dc.identifier.uri | https://doi.org/10.1111/cgf.13794 | |
| dc.identifier.uri | https://diglib.eg.org:443/handle/10.1111/cgf13794 | |
| dc.publisher | The Eurographics Association and John Wiley & Sons Ltd. | en_US |
| dc.title | Unsupervised Cycle-consistent Deformation for Shape Matching | en_US |